Skip to main content

Full text of "Celtic studies from the German"

See other formats

Digitized by the Internet Archive 
in 2013 


















P R E F A C E . 

The history of the variation of opinion about the Celtic lan- 
guages would make a curious and instructive chapter of literary 
history. Their relationships with other languages, like those 
of the peoples who spoke them with other branches of the 
human race, depended rather upon the dictates of passion than 
of reason. There was indeed but little room in most cases for 
the exercise of the reason, because those who theorized about the 
Celtic languages were generally wholly ignorant of them, or, 
at least, knew them very imperfectly, and in their most modern 
and corrupt forms. The rudest tongue is dear to those whose 
first thoughts were expressed in it. The pride which the Irish 
or Welsh take in their language is legitimate, and the exagger- 
ated estimate which they may sometimes form of the beauties 
and powers of their respective dialects can readily be pardoned. 
But the same indulgence cannot be extended to writers who 
contribute to bring science into discredit, and contempt upon 
the language and literature of a people, and therefore upon the 
people themselves, by fanciful and baseless speculations. It 
matters not whether, like Vallancey's, these speculations tended 
to exalt the Celtic language, or, like Pinkerton's, to degrade 
it : both are injurious to the growth of true learning. Indeed, 
the former are the worse, because passages like the following, 
written by Mr. Pinkerton, could only degrade the author: " The 
mythology of the Celtas (which is yet to be discovered !) re- 
sembled in all probability that of the Hottentots or others, the 
rudest savages, as the Celtae anciently were, and are little better 
at present, being incapable of any progress in society". I have 
called up the literary shade of Pinkerton from the oblivion into 
which he has sunk, not because these old opinions are now of 


iv Preface. 

much consequence of themselves, but because they show one of 
the extremes of opinion once held regarding the affinities of 
the Celtic language. This kind of literature now very rarely 
disgraces comparative philology, but, as may be expected in a 
subject like ethnology, which, as yet, scarcely deserves the name 
of a science, and in which mere assertion too frequently usurps 
the place of inductive hypotheses, it still constitutes, if not an 
important, at least a very popular element. 

There is scarcely a language in the world between which and 
the Celtic some one has not attempted to prove a connection ; or, 
to speak more precisely, its chief existing dialect, the Irish. The 
disciples of the Pinkerton school were, of course, desirous that 
its affinities should be with the languages of the inferior races, 
and accordingly one found a great similarity between it and the 
tongue of the Jaloffs, on the coast of Africa ; another found that 
it was a distant cousin of that of the Leni Lenappe, a great 
family of American Indians, who formerly possessed the region 
of the Susquehannah. Others, again, found its true relations in 
the Lappish, the Ostyak, the Tungus dialects, and other tongues 
of North Siberia. On the other hand, the admirers of the Celtic 
tongue endeavoured to establish what, at one time, was con- 
sidered the noblest of origins, a Hebrew descent. This Semitic 
relationship was, no doubt, suggested by the traditions of an 
eastern origin, which pervade the Irish chronicles. As every 
ethnological puzzle was attempted to be solved by means of the 
Ten lost Tribes of Israel, it was of course suggested that the 
Irish were descended from them ; the favourite Semitic ancestors 
of the Celts of the west were, however, not the Israelites, but 
their cousin-germans the Phenicians ; as p is always represented 
in the Irish by/, the bearla fene was the lingua punica: and then 
was not the bdlltaine of May-eve a remnant of the worship of 
Baal ? Carthage was founded by the Phenicians ; the Carthagi- 
nians must, therefore, have been cousins of the Irish, and, conse- 
quently, the fragments of their language preserved in the 
Psenulus of Plautus may be interpreted through the Irish ; and 
so they were. But Sir W. Betham left Vallancey a great way 
behind, when he found that the affinities between the Irish and 
the Hebrew were often so close that he could not detect closer 
between the Irish and Welsh ! 

Preface. v 

There seems to have always existed among writers on 
languages a belief in the great antiquity of the Celtic tongues, 
— that they were much more ancient than most other Eruopean 
languages ; and under this impression is was suggested that the 
Greek, Latin, and even the Sanskrit tongues were derived from 
them, or rather from a primitive Celtic mother-tongue. If even a 
fourth or fifth cousinship could not be permitted with the Greek or 
the Gothic, how could it be tolerated that Celtic should be made 
the progenitor of them all ? Accordingly, such pretensions were 
thus summarily dismissed by a writer who, whatever may have 
been his pretensions as an Orientalist, seems to have had no 
claim to be considered a Celtic scholar, except perhaps that of 
having a Gaedhelic name. " The Celtic, therefore, when divested 
of all words which have been introduced into it by conquest and 
religion, is a perfectly original language; but the originalities 
incontrovertibly prove that neither Greek, Latin, or the Teu- 
tonic dialects, nor Arabic, Persian, or Sanskrit were derived from 
the Celtic, since these languages have not any affinity whatever 
with that tongue" / The tradition which brought the Milesian 
Irish through Spain in their journey from the East, suggested 
an affinity with the Basques and Gascons, which some persons 
have stated to be so close, that an educated Irish-speaking man 
would be able to hold a dialogue with a Basque peasant speaking 
the Escaldunac. There is, of course, not the shadow of a ground 
for this statement, but Irish and Basque affinities are still confi- 
dently spoken of by English writers who know neither the Irish 
nor the Escaldunac tongue. 

The Escaldunac is not the only tongue, the affinities of which 
are still doubtful or obscure, with which the Celtic languages 
have been connected by Engish writers; for Armenian, and 
Albanian, and even Coptic words have been found in them. 
That some affinity exists between Celtic and the first two is of 
course probable enough, as they are now beginning to be consi- 
dered Indo-European ; but the grounds upon which such affinities 
were assumed were as unscientific as those which connected the 
Irish with Phenician. 

It was only through sources like Pinkerton, Vallancey, Betham, 

* Researches into the origin and affinity of the principal languages of Asia 
and Europe, by Lieut.-CoL Vans Kennedy. London, 1828, p. 85. 

vi Jrrejace. 

and Kennedy, that thirty years ago the scholars of France, Ger- 
many, and other foreign countries could have learned anything of 
the Celtic language or literature of these Islands; and scarcely 
anything was known of the Armoric of Bretagne. What won- 
der then that Malte Brun, F. Schlegel, and others, should have 
adopted the opinion of Pinkerton, that Irish was a peculiar lan- 
guage unconnected with the other European tongues ? The first 
man who had the merit of investigating the problem of the 
affinities of the Celtic was the distinguished ethnologist Dr. 
Pri chard, who in 1832 published a supplement to his Researches 
into the Physical History of Mankind, under the distinct title of 
The Eastern Origin of the Celtic Nations} Before the publica- 
tion of this work, Bopp had published his Sanskrit Grammar, 
and J. Grimm his great German Grammar, works which mark 
an era in the history of comparative philology. Dr. Prichard 
was consequently able to base his inquiries upon the labours of 
these great scholars by whom the true foundation of the science 
has been laid. Although this work is now of very little, if any 
use, it was, considering the time at which it was written, and 
that the author appears to have been only able to use chiefly the 
modern forms of the Welsh, in which the inflexions are to a great 
extent lost, a very meritorious work, and one which will always 
be valuable in a historical point of view, as the first in which 
a true scientific method of investigation was attempted. In this 
work Dr. Prichard endeavoured to prove that the true affinities 
of the Celtic languages were with the Sanskrit, Greek, Latin, 
Gothic, and Slavonian, which were considered to form a family 
derived from a single primitive tongue, and to which the name 
Indo-Germanic was given, and furthermore, that it was a mem- 
ber of that family, which should henceforward more appropriately 
be termed the Indo-European, or, as it seems now destined to be 
called, the Aryan family. 

Soon after the appearance of Dr. Prichard's work, and, no 
doubt, owing in a measure to it, the Celtic languages began to 
attract the attention of Continental scholars. Comparative philo- 
logy had now grown into a great science, and was vigorously 
cultivated by many ardent labourers. Between 1837 and 1840, 

b A reprint containing much additional matter, but altogether of an ethnolo- 
gical character, by the editor, Dr. Latham, was published in 1857. 

Preface. vii 

three important works on Celtic philology appeared. The first 
was l)e VAffi.nite des Langues Celtiques avec le Sanskrit, by 
Adolph Pictet (Paris, 1837). The author, who had long 
devoted his attention to the subject of Irish antiquities, having 
published, in 1824, his work on Du Culte des Cabires chez les 
Anciens Irlandais, made the Irish the basis of his study. This 
work still retains its value, and its author is still an ardent and 
respected labourer in the same field. The second work was Die 
Celtischen Spraclien of Bopp, which was published at Berlin in 
1839. This work, which contains several important discoveries, 
may be looked upon as a supplement to bis great work, the 
Comparative Grammar, which did not include the Celtic. The 
third work was the Celtica of Dr. Diefenbach, which was pub- 
lished at Stuttgart in .1839 and 1840. Although this work is 
rather ethnological than philological, yet, as the first part was the 
earliest attempt to bring together the numerous Celtic words, or, 
at least, those which are presumed to be so, that are scattered 
through the works of Greek and Roman authors, and determine 
their comparative etymological relationships with different lan- 
guages, it must always be regarded as one of the classics of 
Celtic philology. 

The honour of having done in a great measure for the Celtic 
dialects what J. Grimm did for the Germanic ones by his cele- 
brated grammar, and of having thus established the basis by 
which the Indo-European character of those dialects could be 
subsequently rigorously established, was however reserved for J. 
Kaspar Zeuss. After thirteen years of labour, he unexpectedly 
presented to the world in 1853 his Grammatica Celtica, written 
in Latin, a monument at once of his genius and of his unex- 
ampled perseverance. In this great work he has left us the 
materials by which we may clearly establish that the Celtic lan- 
guages are pure Indo-European tongues without any admixture 
of heterogeneous foreign elements, and consequently that they 
are members of the family in the same sense that Latin or Gothic 
is. That the labours of his predecessors had not definitely settled 
the latter point, or at least had not brought conviction to the 
minds of many English ethnologists, is very evident from the 
following observation of Dr. Latham : "i relationship was mis- 
taken for the relation. The previous tongues were (say) second 

viii Preface. 

cousins. The Celtic was a fourth or fifth. What was the result ? 
not that a new second cousin was found, but that the family 
circle was enlarged". — Man and his Migrations, p. 87. 

It is right to state that the writer in question does not seem to 
have been influenced in his opinion by the publication of the 
Grammatica Celtica. The passage above quoted was written in 
1851. Here is what he says in 1857: "The real condition, 
however, in which Prichard left the question was this, viz., 
that if the value of the class called Indo-European was to be 
raised by any fresh additions, the Keltic group of languages 
should form either the part or the whole of such additions. 
More than this I cannot find in his paper; more than this I 
cannot find in either Bopp's or Pictet's ; more than this I cannot 
find anywhere. By which I mean that I nowhere find evidence 
upon either of the two following questions : 1 st, That the Kelt (or 
indeed any other language) can be made Indo-European without 
raising the value of the term. 2nd, That any good is effected 
by so raising it. 

" If the writers in question expressed themselves to the fact that 
the tongues in question were absolutely Indo-European, or (still 
more) if they derived them from the East, they left omissions in 
their argument which, to say the least, were illegithnate".— 
PricharaVs Celtic Nations, by Latham, p. 356. 

Dr. Latham, to be sure, seems to attach very little importance 
to the labours of comparative philologists of the German school ; 
for he does not believe in the method of analysis by letter- 
changes. He says, " Whether the clever manipulation of letter- 
changes has, by enabling men to go wrong according to system, 
done as much harm as it is destined to do, is doubtful. It is 
pretty certain that it has done almost all the good of which it is 
capable. For all useful purposes, Prichard used it, the results 
being what we have seen. It is not, then, from this quarter that 
any advancement of Kelt ethnology is to be expected" — ibid., 
p. 382. If the instrument of research in comparative philology 
be not the use of the laws of letter-changes, what is it? Dr. 
Latham does not tell us, at least he does not do so in the follow- 
ing passage : " An improved logic, and a greater sobriety of idea, 
combined with a great breadth of view, are the real desiderata, 
at least for the settlement of the more general questions" — ibid,, 



p. 382. These are desiderata in all scientific inquiries, but they 
do not constitute the method of research of a science. Either 
the changes which the words of any given language undergo 
when that language branches into dialects or distinct languages, 
are arbitrary, or follow regular laws. If the former, the relations 
of languages can only be guessed at from the accidental resem- 
blance which words may offer when placed at random in parallel 
columns; in this case there can be no science of comparative 
philology. If the latter, the first problem for the philologist is to 
determine the phonetic laws of each language ; and no dependance 
can be placed upon any conclusions which may be drawn from 
researches made upon languages, the phonetic laws of which are 
not accurately known. These laws can only be determined by 
careful induction from many and varied researches. Even were 
the phonetic laws of a whole family of languages accurately 
known, it does not necessarily follow that every one could use 
them correctly. As in every other branch of science, a true 
instrument may be wrongly or unskilfully used. No one objects 
to mathematics as an instrument of investigation in physical 
science, because, having been wrongly used, it has sometimes 
led to erroneous results. For the same reason, the mistakes 
made by Leo about the Malberg glosses upon a copy of the Lex 
Salica, or Holtzmann's astounding conclusion that the Gauls 
were Germans and that both were Celts, is no proof against 
the doctrine that a correct etymology can only be arrived at by 
means of a study of the letter- changes. In the hands of Bopp 
and of his school, comparative philology, founded upon a judi- 
cious use of letter -changes, has been raised to the rank of an in- 
ductive science. But this does not imply that Bopp never made 
a wrong induction or proposed a false hypothesis. In comparative 
philology, as in all other sciences, no hypothesis, however logically 
established, can be wholly true; the proportion of error in it 
will, among other things, depend on the state of development of 
the science, and on the greater or lesser generality of the hypo- 
thesis itself — that is, on the greater or lesser number of phe- 
nomena embraced by it. 

This brings me to a more general objection which is raised, 
not merely to comparative philology, but to all sciences — namely, 
that its hypotheses are continually changing. To make this 

x Preface. 

objection, or such an one as has been made to letter-changes, of 
an illegitimate use having been made of its methods, is to 
mistake the scaffolding, by means of which an edifice is erected, 
for the permanent structure itself. If a little more attention were 
bestowed upon the historical development of different branches of 
science, the mistake would not be so frequently made. We should 
then learn what a large amount of scaffolding and useless mate- 
rials are cast aside in the course of a single century's growth- 
scaffolding and materials which may, perhaps, have formed the 
sole subject of that century's intellectual strife. 

Once the Celtic tongues were proved to be Aryan, the detailed 
study of their grammar, from a comparative philological point of 
view, became a necessity in connection with the comparative 
grammar of the whole family. In 1856 a special journal was 
established in Germany, called Beitrage zur vergleichenden Sprach- 
forsc7iung, devoted to the Aryan, Celtic, and Slavonian languages, 
edited by Drs. Kuhn and Schleicher, as a kind of supplement 
to the well-known Zeitsehrift fur vergleichende Sprachkunde, 
founded by Drs. Th. Aufrecht and Ad. Kuhn, and now edited 
by Dr. Kuhn alone, the domain of which is the Germanic, 
Greek, and Latin. Of the Beitrage, a volume consisting of four 
parts appears every two years; three volumes have already 
been published. It is eminently entitled to the support of 
all persons interested in the advancement of Celtic philology, 
and no public library in Ireland, Scotland, or Wales, at 
least, should be without a copy. Besides the papers published 
in this repertory, there is now quite a Celtic philological literature, 
of which I shall only mention a few of the most important works, 
namely, the remarkable book of Gliick, about the Celtic names 
which occur in Caesar (Die bei C. J. Ccesar vorkommenden kel- 
tischen Namen, in ihrer Echteit festgestellt und erldutert von Cr. W. 
Gliick, Munich, 1857); the Ethnogenie Gauloise of the Baron 
Belloguet, which contains a Gaulish glossary, and a collection of 
Gaulish inscriptions ; the Monuments des Anciens Idiomes Gau- 
lois, par H. Monin, Ancien eleve de l'Ecole Normale. Paris, 
1861 ; and the Origines Europceae — Die Alien Volker Europas 
mit ihren Sippen und Nachbarn: Studien von Lorenz Diefen- 
bach. Frankfurt a. M., 1861. 

Among the Celtic papers which appeared in the Beitrage 

Preface. xi 

were a remarkable series entitled Celtic Studies, by Dr. Hermann 
Ebel, the separate titles of each being: — 1. Loss of p in Celtic 
(vol. L, p. 307); 2. Some prepositions (ibid., 311); 3. The 
pronoun som, sera (ibid., 313); 4. Declension (ibid., 155; No. 4 
appeared before the others) ; 5. The so-called prosthetic n (vol. 
II., p. 64); 6. Addenda to Declension (ibid., 67); 7. The gra- 
dation (ibid., p. 78); 8. Phonology (ibid., p. 80). Besides 
these there is a paper entitled Celtic, Greek, Latin, the sub- 
ject of which is the position of the Celtic languages in the Indo- 
European family, and a still more important and elaborate one 
on the same subject entitled the Position of the Celtic. Of the 
Celtic Studies the most important is the paper No. 4, on Declen- 
sion ; it is indeed nearly equal in length to all the others put 
together. Nos. 5, 6, and 7 may be looked upon as supplements 
to No. 4. The object of these papers on declension was to 
determine, according to the principles of the Boppian School, 
the kinds of stems which belonged to the several series of each 
order of declension, according to the classification of Zeuss, and 
attempt from this to determine the case-endings antecedent to 
the oldest forms known, and thus determine the various changes 
which they underwent from the primitive or mother-tongue 
of the family. 

I felt that papers of this kind ought to be brought under the 
notice of Celtic scholars, and especially of Irish scholars, and 
I accordingly undertook to translate the papers on declension 
for the Atlantis. When the translation was complete, I found 
that by itself it would be practically unintelligible to the 
majority of those for whom it was written. Zeuss has the repu- 
tation of being very difficult to be understood, and with equal 
truth the same may be said of Dr. Ebel ; for in the first place 
his papers presuppose a knowledge of the Grammatica Celtica, 
and in the second place because, like the German philologists 
generally, his style is extremely condensed. There is a third 
difficulty, which is, however, a local one. Comparative philology 
is not very much studied in Great Britain or Ireland, and 
although Bopp's great comparative grammar has been translated, 
yet scholars are not in these countries very familiar with the 
method of analysis of the Boppian school. Irish scholars, 
likewise, with very few exceptions, have not hitherto turned 

xii Preface. 

their attention in this direction. Perhaps this is the less to be 
regretted in the case of those who have heretofore devoted 
themselves to the study of the ancient language, literature, and 
historical monuments of Ireland, because, had the object of their 
labours been the mere abstract study of the Irish language, we 
should perhaps not have obtained the great results in a national 
point of view, which those labours have yielded. There is, 
perhaps, no country in Europe, in which in the same space of 
time and under a similar amount of difficulty, so much has been 
done, in about twenty-five or thirty years, for the collection, 
preservation, and publication of the records of its ancient history, 
as in Ireland. So, also, it would be difficult to rival in 
patient and conscientious work and solid learning such men 
as Petrie, O'Curry, O'Donovan, Todd, and Reeves, to speak 
only of those who have occupied themselves with the earlier 
periods of Irish history and archaeojogy. The period has now, 
however, arrived, when the cultivation of Comparative Philo- 
logy, besides its own intrinsic worth, would confer important 
advantages upon Irish literature, and very greatly facilitate the 
study of the ancient MSS. I thus ran the risk of labouring in 
vain, and of missing the opportunity of stimulating some of our 
young scholars to enter, and earn for themselves a name in a field 
of study which is so peculiarly their own, and for the cultivation 
of which they possess so many advantages. Under these cir- 
cumstances, I had no alternative but to prepare an explanatory 
introduction — to venture in fact upon the hazardous undertaking 
of becoming, without any special qualification, the interpreter of 
the German School of comparative philology. 

My first idea was to make an introduction of two chapters ; the 
first to contain an explanation of the nature of roots and stems, the 
formation of stems and their classification, and of derivation and 
composition as distinguished from stems. In the second chapter I 
proposed to give a summary of the case-endings of nouns in the 
several Indo-European languages, in order to afford the student 
an opportunity of comparing the Irish forms with those of the 
other members of the family. As the limits which a periodical 
necessarily imposes were exceeded by the first chapter, which 
was of course the most important for my purposes, I was unable 
to add the chapter on the case-endings. For the same reason, 

Preface. xiii 

as well as on account of pressure of other occupations, I was only 
able to publish, in No. V. of the Atlantis, Nos. 4, 5, and part 
of 6, of the Celtic Studies connected with declension. It has been 
stated above that Dr. Ebel's papers are based upon the Grammatica 
Celtica. To study them profitably, indeed to do so at all, the reader 
must have before him the part of that work on declension. As 
many of those into whose hands the Atlantis was likely to have 
come, may not have had an opportunity of consulting that book, 
I thought it desirable to add in the form of an appendix, a trans- 
lation of the part just alluded to; some of the shortest passages 
in other parts of the book referred to by Dr. Ebel were likewise 
translated, and placed among the foot notes. As the paper 
on the Position of the Celtic possesses interest for a wider circle 
of readers than those on declension, I translated it also, and 
published it in No. VI. of the Atlantis. 

Some friends having suggested that it would be desirable to 
have separate copies of these papers printed before the type of 
the Atlantis was distributed, I thought it a favourable opportu- 
nity to add the Studies omitted through want of space, namely, 
on the Celtic Dual, on the Degrees of Comparison, and an ex- 
tremely important one, 9. Zur Lautlehre, which had been in the 
meantime published in the first part of the third volume of the 
Beitrdge; I have likewise added the chapter on Case-Endings. 
I also took advantage of this opportunity to considerably modify 
the first chapter in several parts, with a view of more clearly 
distinguishing the different kinds of stems, and marking the 
difference between stem-formation and derivation. Although 
Dr. Ebel does not place his paper on the Position of the Celtic 
among his Celtic Studies , I thought it more convenient to do so, 
to avoid the necessity of a long title. I have also put all the 
papers on Declension together as a chapter divided into sec- 
tions, the shorter papers forming in every case a distinct section. 

As it may add to the value of the paper on the Position of the 
Celtic, to give a brief analysis of the discussion out of which it 
arose, I will give here the substance of the note with which I 
prefaced it in the Atlantis. 

So soon as the Celtic was firmly established as a branch of 
the Indo-European family of languages, the next question to be 
determined was its position with respect to the other branches of 

xiv Preface. 

the family. Tlie general opinion at one time was, that the Celtic 
branch first separated from the parent stem. To this early sepa- 
ration was attributed its apparent deviation from the family type, 
above all, the mutilation and partial loss of its inflexions, which 
is found even in the oldest Irish. In an admirable article, pub- 
lished in the seventh volume of the Zeitschrift fur vergleichende 
Sprac7iforschung, Dr. Lottner endeavoured to show, that no 
special relationship could be scientifically established between 
the Hellenic and Italic branches of the Indo-European family, a 
doctrine which must appear heretical to most classical scholars. 
In discussing this subject he had formed the opinion that the 
Celts, Germans, and Lito- Slavonians had lived together as one 
people, and from them the Celts first separated, and then the 
Germans. In a short paper, entitled " Celtisch, GriechiscJi, 
Lateinisc7" (Beitr. I. 429), Dr. Ebel discussed the position of the 
Celtic, and on the whole supported Lottners view of an inti- 
mate relation between the Celtic and German languages. Indeed, 
he appears to have long entertained such an opinion ; for he says, 
in the paper just alluded to: " I cannot deny that already on my 
first acquaintance with Zeuss' Grammatica Celtica, the Celtic 
made an impression on me of an intimate connection with the 
Northern Languages, and that this impression had been conti- 
nually strengthened during my Celtic studies". In the very same 
number of the Beitr age, and immediately following the paper of 
Dr. Ebel, there is a paper by the distinguished philologist Prof. 
Schleicher, entitled, Die Stellung des Celtischen im Indogerma- 
nischen Sprachsta?nme, in which he says : " If in those words of 
Ebel (just quoted) I put Latin, instead of Northern languages, I 
will accurately describe the impression which the study of the 
Celtic made on me". As may be anticipated from this, Prof. 
Schleicher is of opinion, that the Celtic is most nearly connected 
with the Graeco-Latin branch, standing towards those languages 
somewhat in the same relation that the German does to the Slavo- 
Lettish, coming nearer to the Italic (Latin), however, than to 
the Greek. The object of his paper is to bring forward argu- 
ments in support of this Latin relationship, while he left to Ebel 
the task of discovering the agreements between the Celtic and 
the Northern Languages. The paper which is here translated 
is Dr. Ebel's answer to that invitation. Instead, however, of 

Preface. xv 

attempting to determine the agreements in question merely, he 
has taken a wider range, and endeavoured to lay a solid founda- 
tion from which the whole problem of the affinities of the Celtic 
with all the other members of the Indo-European family may 
hereafter be investigated. 

Since the publication of Dr. Ebel's paper, Dr. Lottner has pub- 
lished another under the title of Celtisch-Italisch (Beitr., ii. 309), 
in which, without at all departing from his opinion regarding 
the absence of special affinities between Latin and Greek, he 
has slightly modified his views about the position of the Celtic. 
This chance is due to the light which the Gaulish inscriptions 
have thrown upon the forms of the Old Celtic. These inscrip- 
tions reveal to us words which not only do not yield in antiquity 
of form to those of classic Latin, but even attain, in many respects, 
that of the archaic language of the Romans. They show, beyond 
a doubt, that the inflexions which Irish has retained are older 
than the absence of inflexions in Welsh, and that the wonderful 
phonetic peculiarities of the modern Celtic, the umlaut, aspira- 
tions, the nasals in the Old Celtic, are foreign to it. One inter- 
esting result has followed the investigation of these inscriptions, 
namely, that they give us in part the very forms which' were anti- 
cipated by Dr. Ebel according to the phonetic laws of the later 
Celtic. As I cannot give a translation of the whole of this 
interesting paper, I may, however, state the ethnological deduc- 
tion which he has made. First, as he had already shown in his 
paper published in the Zeitsclwift, the European bough of the 
Indo-European family, after its separation from the Asiatic one, 
foimed a single people, from which the Hellenes (or perhaps the 
Helleno-Phrygians) first separated. The remainder subsequently 
split into two divisions, the South- West and the Northern. The 
former became subdivided into the Italic and Celtic branches, 
while the latter became subdivided into Germans and Slavonians, 
the Slavonians in turn becoming further subdivided into Slaves 
proper and Letts. Of course, much remains to be done before 
this ingenious hypothesis can be looked upon as more than a 
probable explanation, and more than this Dr. Lottner does not 
claim for it. It has much to recommend it, however ; it rationally 
explains the ethnological problem of the present European races, 
and this explanation harmonizes with the ancient Irish tradition 

xvi Preface. 

respecting the Celtic one. Although, genealogical traditions of 
races reaching back into very remote times are not safe materials 
out of which to frame ethnological theories, neither can they be 
altogether disregarded ; and consequently a hypothesis founded 
upon strictly scientific deductions, which, at the same time, 
accords with the popular traditions, may be fairly considered to 
possess many elements of truth. 

It is almost unnecessary to say that an introduction such as that 
which I have prefixed to Dr. Ebel's papers, could, from its nature 
and objects, be to a great extent only a compilation from the works 
of those scholars who are considered to be masters in the science. 
Indeed, I have avoided, wherever I could, introducing any ex- 
amples of my own. In the classification of stems, I have, however, 
ventured to deviate in some degree from that usually followed, 
whether with advantage or not remains to be seen. In an essay 
intended to be merely explanatory of a system, and admittedly 
compiled from the works of those who are authorities upon it, 
it is not necessary to refer to those authorities in every case in 
the text; here, however, it may be useful to mention the chief 
books to which I am indebted for materials. These are : Bopp's 
Vergleichende Grammatik (2nd ed.) ; Grimm's Geschichte der 
Deutschen Sprache; Curtius, Die Bildung der Tempora u?id 
Modi; and Heyses System der Sprachwissenschaft, edited by Dr. 
H. Steinthal. 

A great many notes have been added to the papers by Di\ 
Ebel on Declension, and a considerable number of words added 
to the lists in his paper on the " Position of the Celtic", espe- 
cially to that of the Latin loan-words in Old Celtic. For the 
most of these additions, which are distinguished by being en- 
closed in [ ] , I am indebted to Whitley Stokes, Esq. I also 
take this opportunity to give my best thanks to that distinguished 
scholar, Prof. C. Lottner, from whom Celtic philology has so 
much to expect, for the great pains he took in looking over the 
proof sheets; and also to my friend, John E.Pigot, Esq., without 
whose encouragement the task would never have been under- 

With the \iew of rendering the materials contained in the im- 
portant paper On the Position of the Celtic as serviceable as pos- 
sible in the construction of that great desideratum of Irish lite- 

Preface. xvii 

rature, a dictionary, I have added full Indices Verborum. This 
addition has added much, to the size of the book, but I hope it 
will be found to be a practical contribution to Irish lexico- 
graphic materials. 

In conclusion, I "wish to direct the attention of such of rny 
readers as may not be members of the Irish Archaeological 
and Celtic Society to a work published by that body, which 
contains much that illustrates the subject of the following pages, 
or that is actually supplementary to them, namely, Irish Glosses, 
a mediaeval tract on Irish Declension, icith examples explained in 
Irish, to which is added the Lorica of Gildas, with the Glosses 
thereon, and a selection of Glosses from the Book of Armagh, 
edited by Whitley Stokes, A.B. 

In point of varied learning, skill, and cautious discretion in the 
grammatical analysis, the work is unquestionably the best con- 
tribution to the comparative philology of the Celtic languages 
which has yet appeared in the English language, and may fully 
rank with any similar works by German or French scholars. It 
is at once a valuable and a timely contribution towards the 
materials for making an Irish dictionary, and as such the Archae- 
ological and Celtic Society has well expended its funds in the 
publication of it. 

The most valuable feature of the work in question, so far as 
regards the Celtic Studies of Dr. Ebel, is the large number of 
paradigms of the declension of Irish nouns and adjectives which 
it contains. For the purposes of reference, I think it will be 
useful to enumerate them all. 

Masculine, neuter, and feminine a- and a-stems : noni. sing, cenn, stem cinna 
(masc), p. 39 ; nom. s'mg.forcetal (h), stem forcitala (neut.), p. 51 ; nom. sing, 
masc. mall, an adjectival stem, p. 97 ; nom. sing, rann, stem rannd (fem. ci- 
stern), p. 38 ; nom. sing, clia, a masc. a-stem, p. 45. 

Masculine and feminine la-stems : nom. sing, rannaire, stem rannaria (masc), 
p. 37 ; nom. sing, caile, stem calid (fem.), p. 54 ; nom. sing. masc. nue, an adjec- 
tival /a-stem, p. 97. 

Masculine and neuter z-stems; nom. sing, faith, stem fdthi (masc), p, 36 j 
nom. smg.Jiss, stem Jissi (neut.), p. 117. 

Masculine w-stems : nom. sing, hlih, stem bithu (masc), p. 62. 

Masculine osteins : nom. sing.Jili, stem filiat (masc), p. 36. 

Masculine (/-stem : nom. sing, ri, gen. rig, a masculine g-stem, p. 119. 

Feminine n-stem : nom. sing, talam, stem talaman, p. 48. 

Anr-stems ; nom. sing, cara, stem carat, from carant (masc), p. 65. A para- 

xviii Preface. 

digm of the declension of ainm (n), probably originally an ant-stem, but which 
was in Old Irish a neuter ann-stem, is also given at p. 116. 

Masculine r-stem : nom. sing, athir, stem athar, p. 39. 

C-stems : nom. sing, cathir. According to Dr. Ebel (see p. 94), cathir is an r- 
stem, taking the determinative suffix c, but Mr. Stokes considers it to be a c-stem, 
p. 38. 

Anomalous nouns : nom. sing, ben, all the singular and plural forms of which 
are given, p. 121. 

At p. 45 a paradigm of the declension of tlie article is also 
given. What renders these paradigms the more valuable is, 
that in almost every case the forms of the dual number are also 
given. As several of the words declined by Zeuss and Dr. Ebel 
are also declined by Mr. Stokes, the corresponding paradigms of 
each writer may be instructively compared. 

Dr. Ebel's papers are frequently referred to in Mr. Stokes's 
book; and as each may be said to, in a measure, supplement 
the other, the almost simultaneous appearance of the following 
translation of the Celtic Studies, and of the admirably edited 
book in question, may be deemed a fortunate coincidence. I 
hope, also, that the introduction which I was obliged to prefix 
to the papers of Dr. Ebel may likewise enable a larger circle of 
readers to appreciate the importance of Mr. Stokes's contribution 
towards our more perfect knowledge of the language of Ancient 


Chapter I. — Ox simple v, op.d-formation : roots, stejis, and derivatives. 

§ 1. Of Roots and Root-forms. Nature of a root as a phonetic symbol, p. 3. 
The permutations or letter changes which take place in the same word 
in different dialects of the same language, or in different languages of 
the same family, not arbitrary, p. 3. Example of the law of transposi- 
tion orprovection of sounds (called by Grimm Lautverschiebung), p. 3. 
The analysis of words which have had a common origin does not give 
roots, but only root-forms, p. 4. Root-forms of the same root often 
very dissimilar : examples, p. 4. Objects of modern comparative ety- 
mology, p. 4. Modern classification of the kinds of words of which 
rational speech is composed, p. 4. Corporal and formal, or formational 
words, p. 5. Corporal and formational roots, p. 5. Of the composition 
of roots in words : 1. Parathesis; 2. Agglutination; 3. Amalgamation, 
p. 6. Classification of languages according to their relative degree oi 
composition, p. 6. Bopp's classification somewhat different from that 
here given, p. 6. The amalgamating languages the most complex 
are represented by the Indo-European or Aryan family of languages 
such languages contain no uninflected roots in their primitive state 
but many naked roots occur in their modern derivatives, pp. 6-7. 
Roots preserve their identity ; exception to this in the process called 
Root Variation ; this process gives rise to affiliated roots, one of which 
is primary to the others ; all large roots are secondary ; example?, p. 7. 

§ 2. Of Elementary Word-formation and Inflexion. "Word-formation or 
Word-building ; elementary words must be further modified by inflexion ; 
the process of elementary word-formation and inflexion the same ; what 
those are : — 1. internal phonetic change ; 2. addition of phonetic material, 
p. 7-8. Vowel change a predominant process of this kind in the Semitic 
languages ; only appears as ablaut in the Indo-European tongues ; defini- 
tion oi ablaut ; chief agent of word-formation in the Teutonic languages, 
p. 8. Of umlaut ; it acquires inflexional signification : exanrples, p. 8-9. 
Of Breaking or Fracture, p. 9 (note on the German nomenclature em- 
ployed, note 4, page 8). Progressive assimilation in the Finno-Tata- 
rian family, a kind of umlaut ; the rule " caol le caoV in Irish, and the 
weakening of the root vowel in Latin by the vowel of a prefix, may 
be considered as progressive assimilation, p. 9-10. Affixes the chief agent 
of word-formation in Indo-European tongues ; the affixes used in w r ord- 
formation and inflexion, and in composition ; some of the affixes used 
in word-formation and inflexion traceable to independent words ; — ex- 
amples; generalization of this result; word-formation and inflexion 
were originally synthesis of independent roots ;— this view is the basis 
of the agglutination theory ; the theory now generally admitted ; some, 
however, modify it so as to admit two kinds of affixes, p. 10-11. 

xx Contents. 

§ 3. Of the stem, a form intermediate between the root and the word ; dis- 
tinction between roots and stems ; stems are of many kinds ; hence 
there are four categories of phonetic forms, p. 11-12. Word-formation 
consists of stem-formation and derivation; phonetic methods of 
primary or pure stem-formation. — I. Modification of the root vowel : — 
by (a) ablaut proper; (b) obscuration; (c) strengthening — either (1) 
hy lengthening a short vowel, or (2) by gunation and diphthongation, 
(note explanatory of those terms). II. Consonantal strengthening of 
the root by — (a) duplication of final consonant ; (b) affixation of a 
foreign mute consonant ; (c) affixation or intercalation of a nasal — by 
(1) nasalizing an internal vowel, (2) affixation of the nasal in the 
auslaut, either after vowels or after consonants, or (3) affixation of a 
whole syllable accompanied by nasalization ; or (d) reduplication ; — 
examples, p. 12-14. Exceptional cases of primary stem formation; 
e. g. an intensive introduced into the root ; other classes of true stems ; 
occasional difficulty in distinguishing stems from roots, and simple 
words from stems, p. 14. 

§ 1. Of noun-formation and secondary stems. Separation of grammatical 
categories by the addition of signs to stems ; signs by which a stem 
becomes a verb ; sign of noun-formation (to which the present analysis 
is confined) ; distinction of gender ; nominative sign — masc, fern., and 
neuter; Bopp's view as to origin of neutral adj., ending-ata ,• relative 
degree of preservation of the nominative sign s in different languages 7 
p. 14-16. Introduction of a vowel between the stem and ending ; the 
declension vowel (note 7 on the Finnish two-syllabled Stems), p. 16, 
Cases of stems formed by a whole syllable — -derivative Stems, p. 16-17. 
Two-fold classification of stems according to auslaut and relation to the 
grammatical ending : I. vocalic stems — 1. pure stems ; 2. middle forms 
produced by affixing a declension vowel; II. consonantal stems — 1. 
pure stems ; 2. middle forms produced by affixing a syllable ending 
^consonantally, p. 16. Table of the classification of true stems, and of 
derivative stems, p. 16-17. 

§ 5. Of vocalic stems. [I.] Pure stems; examples: Greek, Latin, Gothic; 
Gothic nouns properly belong to second class, the Teutonic languages 
having no pure stems (but see note 12, p. 28, for some probable excep- 
tions), p. 18-19. [II.] Middle forms ending vocally. Division of all 
vocalic stems into a- stems, i- stems, and u- stems, p. 19. (1) I. Stems, 
Greek, Latin, Gothic ; peculiarities of adjectives derived directly from 
stems, p. 19-20. (2) A- Stems. In Greek and Latin : 1. where the 
primitive a is preserved or changed into e, — (a) stems with primitive 
short a, (/3) stems with long a or e ; 2. where the primitive a is changed 
into o in Greek, and u in Latin. A- stems with primitive a short ; 
(examples). Gothic a- stems ; (examples) ; peculiarities of Gothic ad- 
jectives — examples, p. 20-22. [III.] yd-, or IA- stems. Peculiarities 
in the Gothic, the 0. H. German, M. H. German, and N. H. German 
adjective forms of ya- stems — examples from the Latin and Gothic, p. 
22-24. [IV.] Consonantal stems changed into vocalic (a- and i-) 
stems. Examples from the Latin and the Greek. Apparently irre- 
gular Greek stems. Analogy of certain Gothic nouns with the Latin 
All these probably consonantal n- stems, (note 10, expressing a 
doubt as to the hypothesis in respect to the Greek stems, and 
giving a different explanation of Ahrens, p. 2i). Tables of ex- 
amples. Gothic feminine nouns ending in -ei; probably mutilated 
n- stems; (examples). Declension, in the Germanic (though not 
in the Latin or Greek) languages, affected by the dropping of 
the n. Examples. Peculiarity of the weak declension, in adding 
the n. The weak adjective declension ; (examples). Examples of 
vowel endings distinguishing genders. The s in "sanguis", p. 
24-26. [V.]w- stems. Greek, Latin, Gothic. Peculiarities of adjec- 
tives, p. 26-27 

Contents, xxi 

§ 6. Of consonantal stems. [I.] Pure stems. (1) S- stems ; Greek and Latin. 
(2) Stems with sonant auslauts (semi-vowels : m, I, ?i, r, ng) ; Greek 
and Latin. (3) Stems with medial auslauts ; Greek and Latin. (4) 
Stems "with tenuis auslauts ; Greek and Latin. (5) Stems with 
aspirated mute auslauts ; Greek. Gothic pure stems belong rather to 
middle forms ; examples (note 12 already referred to giving examples of 
some which are perhaps exceptions), p. 27-28. [II.] Consonantal 
middle forms. The " thema'', distinguished from the true stem form. 
(1) S- stems — Greek and Latin. (2) Stems with sonant auslauts. 
L- stems. N- stems, R- stems, (Latin, Greek, and Gothic). (3) Stems 
with medial auslauts. (4) Stems with tenuis auslauts (examples 
from Greek and Latin), p. 28-31. 

§ 7. Of Derivation. Affixes : 1. single letters or syllables, not traceable to 
independent words; 2. syllabic affixes, once independent words, but 
modified so as to have lost that character. Difficulty of distinguishing 
stem-formation from derivation. " Derivative Stems'' (Greek examples). 
Examples showing the formation of words by the process of Derivation, 
pp. 31-33. 

§ 8. Of Composition. Definition. Peculiar power of composition in the 
Sanskrit and the Greek; in the Germanic languages. Composition 
distinguished into: 1. synthetical; and 2. parathetical. Particle com- 
position to be classed as parathetical. Introduction of a copulative 
vowel in the older languages ; Greek and O. H. German ; Modern 
German, and English (examples). Combination sometimes accom- 
panied by phonetic changes in one or both the constituents (examples 
in the Latin). Relation of the constituents of compound words (ex- 
amples), pp. 33-34. 

Chapter II. — On the case-exdincs of nouns in the chief Indo-Euro- 

§ 1. The Accusative Singular. — (As to the nominative, refer to Ch. 1., § 4). 
The sign of the accusative: m, in Sanskrit, Zend, Latin ; n, in Greek, 
Lithuanian, and Old Prussian. Primitive sign probably m. In the 
Latin: m affixed, 1. directly to vocalic stems; 2. Avith intercalated co- 
pulative to consonantal stems. Transformation of Sanskrit m into 
nasal ??, by " anusvdra" ; (note on " anusvara"). Observations as to 
the Lithuanian, Latin, Oscan. Nominatives in modern languages 
(Italian, Portuguese, Spanish), derived from the mutilated Latin ac- 
cusative. Greek and Latin declensions compared. The Greek affix v. 
Exceptions (stems in w and £v). In the Gothic the sign of the accusa- 
tive lost, pp. 34-36. 

§ 2. The Genitive Singular. — The Sanskrit and Zend, p. 3G. The Latin. 
The genitive in -ius, pp. 36-37. The Oscan, p. 37. In the Greek; 
gen. sing, formed by : (a) affixing g to fern. 1st decl. ; (b) affixing c 
with a copulative ; (c) where the noun does not form the gen. in g , 
pp. 37-38. The Gothic. The Germanic languages, pp. 38-39. The 
Lithuanian and Slavonian, p. 39. 

§ 3. The Dative, Locative, and Instrumental, Singular. The Sanskrit and 
Zend, p. 39. The Latin. The Oscan, p. 40. The Greek, p. 40. The 
terminations of adjectives (masc. and neut., and fern.), in the Gothic, 
O. H. German, and M. H. German, compared; (Paradigm), p. 41. 
The Lithuanian and Slavonian, pp. 41-42. 

§ 4. The Ablative Singular. Peculiarities in the Sanskrit, Zend, and Latin. 
Ablative sign recognizable in adverbs, p. 42. 

§ 5. The Dual. Peculiar forms only found in Sanskrit, Zend, Greek, Sla- 
vonian, and Lithuanian. (Note : reference to Ebel's observations on 
the relics of the Dual in Irish). Traces in Latin, pp. 43-44. 

§ 6. The Nominative and Vocative Plural. Comparison of the Sanskrit, 
Zend, Old Latin, and Greek. Endings in i. Neuters in a. Plural 
endings in -as; opinions of philologists, pp. 44-45. The Gothic: 

xxii Contents. 

(Paradigms of the strong and the weak declension). Comparison of 
the Gothic with the Old and the Middle High German ; (Paradigm). 
Formation of the plural in the modern languages, pp. 45-46. The Li- 
thuanian and Slavonian, p. 46. 

§ 7. The Accusative Plural. The Sanskrit, Zend, Greek, Oscan, and Latin 
compared, p. 47. The Gothic and O. H. German, p. 47. The Gothic 
-ns. Examples, p. 48. Parallel in Greek, p. 48i 

§ 8. The Genitive Plural. The Sanskrit (-dm), and Zend (-anm), p. 48. 
The Latin (-rum), pp. 48-49. The Greek (-wv), p. 49. The Gothic. 
The M. and N. H. German, p. 49. The Lithuanian and Slavonian, p. 50. 

§ 9. The Dative, Locative, Instrumental, and Ablative Plural. The San- 
skrit and Zend, p. 50. Two forms in the Greek and Latin, pp. 50-51. 
The Gothic, 0. and M. H. German, pp. 51-52. The Lithuanian and 
Slavonian, p. 52. 

Paradigm of all the case-endings of nouns in the chief Indo-European Lan- 
guages. (The Sanskrit; Zend; Latin; Oscan; Umbriau; Greek; 
Gothic ; O. H. German ; Lithuanian ; Old Slavonian ; Old Prussian). 
Folding plate facing p. 52. 

Chapter I. — On declension and the degrees of comparison in Irish. 

§ 1. Dopfis View oj the Aspirations and Eclipses in Modern Irish, and the 
modifications which it undergoes through the old Irish forms. Zeuss' de- 
termination of the Old forms of the article ; (notes, giving the passages 
from Zeuss). Bopp's opinions modified as to the t and the h before 
vowels. (Note on dona from donabis, correcting Dr. Ebel's theory re- 
garding the 0. Irish dat. plur.), (note on "aspiration'' — "Infection", 
or " mortification"), (note on the t in the nom. sing. masc. of the Irish 
article), pp. 55-56. Bopp's opinion modified as to the nom. plur. masc, 
p. 57. Discovery of the neuter, and of the accusative cases in Old 
Irish, p. 57. 

§ 2. Stems which belong to the several orders and series of Zeuss. Objec- ' 
tions to views of Pictet and Bopp as to the distribution of the vocalic 
stems, pp. 57-58. Consideration of the words and suffixes which be- 
long to the several classes. Examples, pp. 58-59. Examples of forms 
in cognate languages, pp. 59-60. Adjectives, pp. 60-61. Verbal sub- 
stantives taking the place of the infinitive,' pp. 61-62. Eorms of the 
article, p. 62. 

§ 3. Test afforded by Irish Phonology for determining inductively the Primi- 
tive Forms oj the Celtic Case-endings. Two close points of contact be- 
tween the Irish and the German vocal systems : umlaut of a by i and 
u, and fracture of i and u by a. Examples, pp. 62-63. The vowel of the 
ending determinable by the vowel changes in the stem, p. 63. Table 
of masc. and neut. endings. (I.) ; (examples), pp. 63-64. (Note : table 
of hypothetical endings of masc. and neut.), p. 63. Table of endings 
of fern, a- stems ; (examples), p. 64. Table of endings of masc. stems 
(III.) ; (examples), p. 64. Table comparing auslaut in Old Irish and 
N. H. German, p. 64. How the Gaedhelic has been harder than the 
Gothic ; (examples), p. 65. Explanation of mutilations of the auslaut, 
p. 65. Table : (Primitive period, — Pre-historic period, — Historic pe- 
riod), p. G6. 

§ 4. Declension of Consonantal Stems. Zeuss' five classes. Analysis, p. 66. 
Table of common endings, p. 67. As to the length in the ace. plur. 
comparison with the Greek, the Latin, and the Gothic, p. 67. Obser- 
vations on the several cases. Examples, pp. 67, 69. Table of declen- 
sion of neutral n- stems, (I.), p. 69. Nouns of relationship in -thar, (II.), 
p. 69. Table : (Primitive Period, — Pre-historic Period, — Historic 
Period), p. 70. Addition of a " determinative suffix", ; (but, — note on 

Contents. xxiii 

this theory of Ebel), p. 70. Comparison of Gaedhelic with the Classic 
languages as to the consonant declension of the t-, n-, and r- stems, p. 71. 
§ 5. Declension of Masc. (and neut) a and ia- Stems, Stems included in 
the vocalic declension, p. 71. Inflections of masc. a- stems; compared 
with Sanskrit, etc. ; (examples), pp. 71-72. Anomaly in the neuters, 
preparing for disappearance of the neut. in the Gaedhelic, pp. 72-73. 
As to adjectives, p. 73. 
§ G. Declension of masc. I and U Stems. Consideration of each case, pp. 
71-76. Table of declensions (of U stems and I stems), arranged ac- 
cording to periods (Primitive, Pre-historic, Historic), but without the 
secondary forms, p. 7G. 
§ 7. Declension of Jem. A and I stems. Confusion in their declension; how 
the primitive stem only now to be recognized ; (but, — note on this). 
Examples, pp. 70-77. Hypothesis of Dr. EbeL pp. 77-78 (see also for 
completion, p. 154). Tables of forms of ia- stems compared with 
those actually occurring, p. 79. Other examples of same degeneration 
of original forms, p. 79. Modern Irish losing its inflexions, like the 
Kymric. Examples, pp. 79-80. 
§ 8. The distinction of the plural in the Kymric No inflexions preserved in 
Kymric except distinctions of plural, and this very arbitrarily employed. 
As in the N. H. German: (1) old plural form remaining, and conse- 
quently true inflexion ; (2) stem-ending preserved, dropped in the sing. ; 
(3) a determinative suffix, wholly foreign in place of the ending, p. 81. 
To (1) belong: 1. Kymric plur. without enlings; (examples), p. 81; 
2. plurals in -i; (examples), p. 81 ; 3. plurals in -an and -iav, p. 82. 
To (2) : especially n- stems ; (examples), p. 82. To (3): 1. many plur. 
in -au and -iau in which the ending is foreign to the words — stem 
proper ; (examples) ; 2. most words in -ion or -on ; (examples) ; 3. 
endings -et, -ot, -ieit, -eit, and -ed, -yd, -oed ; (examples), pp. 82-83. 
§ 9. Note on a-, i-, d-, t-, and nt- stems. Gen. in -i and nom. in -as, in a 
stems, found in Ogam inscriptions. Obscuration of a to o at a remote 
period, p. 83. The neutral aill, p. 83. As to Mr. Stokes' corrections, 
recognizing stems in -d, -t, and -nt, in Zeuss' Ordo Posterior, ser. 4, pp. 
83-8-1. As to Dr. Ebel's view of the fern, in Zeuss' Prior, Ser. V., p. 
84 ; fusion of i- and a- stems, p. 86. 
§ 10. On the Celtic Dual. Answer to Mr. Stokes. "Whether the Celtic has 
a dual ; (examples), pp. 85-86. How much of it has been preserved, 
pp. 86-87. Of undoubted dual-forms only the masc. and ace. of sub- 
stantives, and the whole of the cases of the numeral Two, pp. 87-88. 
The in of the article, p. 88. Few dual forms of consonantal stems pre- 
served, p. 88. 
§ 11. On the Article in Modern Irish. Theory as to the article an, p. 88. 
(But — note questioning this theory of Dr. Ebel), p. 89. Observations 
on certain finer influences of neighbouring languages on one another, 
p. 89. 
§ 12. On the so-called Prosthetic N. [" Prosthetic"=" Eclipsing". Theory 
of Mr. Stokes and of Dr. Ebel, p. 90.] Correction of Zeuss' views as to 
this n. Mr. Stokes' examples. Examples of this n as a relic of the 
article, p. 90. Other examples, p. 91. Some spurious prepositions 
recognized as accusative forms, p. 91. The n of ainm-n; previous ob- 
servation of Dr. Ebel corrected in note, p. 91. Supposed three-fold 
preposition do-air-in, p. 92. The n after verbal forms ; examples, p. 92, 
§ 13. On the degrees of comparison. As to the -ns stems. The -a in the 
more ancient, -u in the newer secondary formations, p. 92. Explana- 
tion of these formations, pp. 93-94. 

Chapter II. — On the position of the Celtic 

§ 1. Views regarding the special affinities of the Celtic and words borrowed 
from the Latin. Points of contact between the Celtic and the Italic 

xxiv Contents. 

tongues on one side, and the Teutonic on the other, p. 97. Views of 
Dr. Lottner and Professor Schleicher, p. 97. Celtic closer to Latin 
than to Greek, p. 97. Points of agreement "between the Celtic and 
the Northern tongues, p. 93. "Words in Celtic languages in common 
with or borrowed from the Latin, p. 98. Method of denoting the 
Kymric dialects, p. 98. Glossary of Latin loan-words in Old Celtic, 
pp. 99-107. Words borrowed by the Latin from the Gaulish, and 
later from the British, p. 107. 
§ 2. Glossarial affinities of the Celtic and Classic languages. Glossary of 
words and roots exclusively common to the Celtic and Classic 
languages, pp. 107-109. 
§ 3. Glossarial affinities of the Celtic, Classic, Teutonic, and Lito- Slavonian 
languages. Glossary of words and roots common to the Celtic and 
Classic languages, but also found in the Teutonic, Slavonian, and 
Lithuanian, pp. 109-112. Of certain other roots to be added to this 
list. Examples, pp. 112-113. 
§ 4. Glossarial affinities of the Celtic, Classic, and Teutonic lauguages. 
List of words and roots common to the Celtic, Greek, and Teutonic, p. 
113. List of those common to the Celtic, Latin (or Italic), and 
Teutonic, pp. 113-114. 
§ 5. Glossarial affinities of the Celtic, Teutonic, and Lito- Slavonian 
languages. Mutual borrowing among the languages. Examples of 
borrowed words, pp. 1 14-1 15. List of words and roots common to the 
Celtic, Lithuanian, Slavonian, and Teutonic, pp. 115-116. 
§ 6. Glossarial affinities of the Celtic and Teutonic. List of words and 

roots common to the Celtic and Teutonic, pp. 116-119. 
§ 7. Glossarial affinities of the Celtic and Lito- Slavonian. List of words 
and roots common to the Celtic, Lithuanian, and Slavonian, p. 119. 
Original words in the Celtic, p. 119. Agreement with the Sanskrit 
in nomenclature of the cardinal points, p. 119. Summary of results of 
the foregoing tables, as regards the true relation of the Celtic to 
other European languages, pp. 119-120. 
§ 8. Phonological affinities: — Vocalismus. Study of principles on which to 
judge of an earlier or later separation of tongues yet imperfect. Ex-, 
ample : comparison of treatment of the neuters in the Old Gaedhelic 
and the Hebrew, — the Polish and the Slavonian. Want of a geography 
of sounds ; (note on this subject), pp. 120-121. The elementary de- 
velopement of the vocalismus only to be followed out with clearness 
in the Gothic, p. 121. The Gothic short vowels, a, i, u, p. 121. The 
Latin and Greek, and the Celtic compared, pp. 121-122. Analogous 
vowel changes in Teutonic, Slavonian, and Celtic roots, p. 122. In 
the diphthongal system the Celtic nearest to the Teutonic. Examples, 
p. 122. 
§ 9. Phonological affinities: — Consonantismus. Celtic analogous to Lithua- 
nian and Slavonian in having no aspirate in its older phonetic stage, 
p. 122. Celtic in this contrasts with the Greek, p. 123. Deviations 
from the Teutonic, p. 123. Agreements, p. 123. Changes of secondary 
aspirates into medials, or medial-aspirates, p. 123. Hardening of 
medials in the Celtic and Teutonic, p. 123. The Gaedhelic thicken- 
ing of the n (or mi), p. 123. 
§ 10. Affinities of word-formation. The suffix -tion exclusively Italo-Celtic, 
p. 123. Other suffixes ; (-U ; -id ; -aire ; -ire ; -doit ;) p. 124. The peculiar 
suffix-combination : antat, {-atu, -etu), p. 124. Celtic word-formation 
of a modern character, p. 124. Wider use of K than in the Classic 
languages ; (-acK), p. 124. 
§ 11. Affinities of declension. Only the Pelasgic languages have fern, a- 
stems. Agreement of Celtic with the northern languages, p. 124. 
Masc. a- stems foreign to the Celtic, p. 125. Few fern, u- stems in 
Celtic, p. 125. Celtic approaches the Classic languages in having 
preserved more pure consonantal stems, but differs from them in treat- 

Contents. xxv 

ment of s- stems, p. 125. As to the ablative, p. 125. As to the 6 in 
the dative plural, p. 125. Want of pronominal declension, p. 125. 
Agreement of gen. sing, and nom. plur. of masc. a- stems in the Old 
Gaedhelic and the Latin, p. 1 25. Other agreements less exclusive, pp. 

§12. Affinities of gradation (cr comparison). Peculiar forms of gradation 
in the Greek, the Latin, the Celtic. Analogy of some Gaedhelic 
forms, p. 126. 

§ 13. Affinities of the pronouns. Celtic peculiarity in giving up the nom. 
sing, of the 1. and 2. person, p. 126. Analogy with the Teutonic in the 
3. person, p. 126. Analogy with forms in the Sanskrit, p. 127. The 
ta ; the ana ; p. 127. 

§ 14. Affinities of conjugation. Peculiar combinations and new formations 
in conjugation. Examples, p. 127. Remarkable analogy with Teutonic 
and Slavonian, p. 127. Paradigm of Old Gaedhelic and Lithuanian — 
ending of the present and preterite, p. 128. The Kymric -st, (2. pers. 
sing, praet.) p. 128. Pictet's view of this -t, p. 128. Distinction of 
the imperfect and perfect in the Slavonian by separate verbs, p. 128. 
Use of the present as a future, p. 129. Peculiar force, in the Teutonic 
and Slavonian, of the particle in composition, p. 129. Analogy in the 
Celtic ; (1) the perfect denoted by a special particle ; (ru-; ro- ; ra-), 
pp. 129-130. Peculiarity of Celtic in use of this particle, p. 130. (2) 
The pres. and fut. changed into the perfect future exactum by this 
particle (ro-), p. 130. (3) The present forms (especially the con- 
junctive pres.) turned into future by it, p. 129. All three uses in the 
Gothic, p. 131. Gaedhelic particle to tenses of incomplete action, 
(nu-, no-), p. 131. Middle position of the Celtic, between the Italic 
and Greek, and the Teutonic and Lito- Slavonian, p. 131. Other points 
of contact to be sought in a comparative syntax of these languages, 
p. 131. 

Chapter III. — On Phonology in Irish. 

§ 2. Necessity of establishing an organic orthography ; and great importance 
of a comparison of the modern Irish forms for the purpose. Schleicher's 
opinion, p. 135. Want of linguistic materials on the continent, p. 
135. Inaccuracy of those published: (examples in Zeuss, O'Dono- 
van, etc.) p. 135. How to attain what is required, p. 135. Com- 
parison wanted between Middle and Modern Irish forms, p. 136. 
Disfigurement and irregularity of Modern Irish forms; (errors of 
Pictet and Bopp), p. 136. Examples, pp. 136-138. Necessity of com- 
parison with newer forms nevertheless ; (error of Zeuss), p. 138. 

§ 2. Vocalismus. The chief difficulty of the Irish phonetic system, p. 138. 
Three kinds of e and o, p. 138. Suggestion of a mode of distinguishing 
them in print, p. 138. Examples, p. 139. Of the a corrupted from 
the o in Old Irish. Examples, pp. 1 39-141. Correction of mistake in 
preceding chapter, (p. 88 ; § II. On the article, etc.) as to the modern 
form of the article an, note, p. 140. 

§ 3. Consonantismus. — Aspiration of mediae after vowels. Important results 
of comparison of the newer forms. Examples, p. 141. Influence of 
the s, p. 141 . Comparison of the modern forms especially necessary to 
deter dine whether the tenuis or media is to be aspirated or not. Ex- 
amples, pp. 141-142. Aspiration of the simple m in Modern Irish. M. 
for bh. Mm (or mb) in Old Irish, deduced from m (in inlaut) in Mo- 
dern Irish. (Note : mm in several examples, compared with the nn 
of the article), p. 142. Mediae after vowels always aspirated in Modern 
Irish ; after consonants not so, except where a vowel dropped out, pp. 
142-143. Mediae assimilated after iquids, p. 143 Observation on the 
so-called Eclipse, p. 143. 


xx vi Contents. 

§ 4. Consonantismus. — Aspiration of tenues after vowels. Tenues when 
aspirated, pp. 143-144. Organic Medias changed into tenues in Old 
Irish, in two ways, p. 144. Comparison of Modern Irish, pp. 144-145. 
Observation as to the so-called eclipse of the tenuis, p. 145. Conclu- 
sion : (Examples), p. 146. 

§ 5. Consonantismus. — Cases which afford occasion for aspiration after a 
preserved or lost vowel: (I.) in inlaid ; (II.) in anlaut; (III.) in syntax. 
Inference of aspiration from the presence of a vowel, p. 146. Examples. 
Three categories. 
(I.) In inlaut, p. 147. Examples from conjugation, p. 147. Confirma- 
tion of Zeuss as to the /, n, s, d, t, th. (Observation as to O'Donovan 
on the Modern Irish Passive and Participles), pp. 147-148. Examples, 
from declension, p. 148. Derivatives in ~te, p. 149. Derivation with 
various suffixes, p. 149. 
(II.) In anlaut. Of the second member in composition, p. 150. Omis- 
sion by Zeuss as to the exceptions to the aspiration rule, p. 150. 
Grimm's observations as to t and d, in Mod. Irish, remaining un- 
changed after liquids, not quite correct. ( u Mactire", explained), p. 
150. Examples of other exceptions, p. 1 50. Explanation of " Dunpe- 
leder" in Zeuss, p. 151. No aspiration following and Mac, in names, 
p. 151. 
(III.) Caution as to use of Mod. Irish in determining laws of anlaut (not 
developed by Zeuss), p. 151. Phonetic changes, how produced in syn- 
tax, p. 152. (1) Original terminations of the article, in the several 
cases. Examples, p. 152. Phonetic laws after the article, p. 153. 
Kule as to Eclipse (O'Donovan), apparently inexplicable, p. 153. Ex- 
planation, however, by comparison of O'Donovan's examples from 
Keating, p. 154. Observations as to confusion of case-endings (p. 78 
et seg.) completed, p. 154. Peculiar use of ace. for nom. in Old Irish, 
p. 154. Confusion in the spoken language ; (examples), p. 154. The 
true ace. in the so-called dat. sing., p. 154. Comparison with Mod. 
Greek, as to loss of dat., p. 155. Example (Table of Declension) of the 
treatment of the anlaut after the article, p. 155. Explanation of so- 
called Eclipse of s after is {in), p. 156. Of the Adjective after the, 
article, p. ) 56. (2) Influence of auslaut on following anlaut between 
adj. and subst., p. 156. Examples in Zeuss few, p. 156. Examples 
for the aspiration of the adj., p. 157. Suppression of aspiration in cer- 
tain cases, p. 157. Transvected nasal, p. 157. Examples for the as- 
piration of the subst., p. 157. Correspondence with what is known of 
the Dual : (O'Molloy and O'Donovan), p. 158. Eclipse after numerals, 
p. 158. (3) Combination between subst. and succeeding genitive 
much weaker, p. 158. Examples of nasal preserved in ace, p. 158. 
Other examples, p. 159. (4) Pronouns : (Examples), p. 159. (5) 
The anlaut after prepositions and other particles, p. 160. (6) Action 
of the verb on the object, p. 161. 

§ 6. Loss of P in Celtic. Preservation of the guttural in the Gaedhelic 
(replaced by the labial in the Greek and the Kymric). Examples, 
p. 161. Primitive p replaced even by c or eh. Examples, p. 161. 
Aversion to p in anlaut. Examples, pp. 161-2. Want in the Celtic 
languages of the prepositions with/?- anlaut in the Sanskrit and other 
cognate languages, p. 162. Assumption by Pictet and Bopp: (exam- 
pies). "Frith". "Fir", p. 162. The Sanskrit " pra" and "pari", p. 
1 63. The prefix " ro",^ " ire",_ pp. 163-4. 
§ 7. Loss of the P in Celtic (continued). Example of loss of p in anlaut in 
" en" (avis), p. 164. Pictet and Pott as to " are", p. 164. Prepositions 
in Old Irish in a double, and even a treble form, p. 164. Examples, p. 
165. Fundamental meaning of " af', p. 165. 




L Series ; II. Ser. ; 
Ser., p. 1 70. 
Plural ; (exam- 




"Translation of the part of Chap. II. of Zeuss' Grammaiica Celtica, referred to 
by Dr. Ebel.] 

A. Declension. Two orders of declension, p. 169. 

First Order.— The " Vocalic", p. 169. 

Declension of Nouns, rnasc. and neut. Paradigm. 

III. Ser., p. 170. 
Declension of nouns Jem. Paradigm. IV. Ser. ; V, 
I. Series. External Inflexion ; nouns in -e, p. 170. 
Singular; (examples in all the cases), p. 170. 
pies in all the cases), p. 171. 
II. Series. Internal Inflexion, p. 171. 

Singular; (examples), p. 171. Plural; (examples), p. 

III. Series. External Inflexion, except dat. sing., p. 172. 

Singular; (examples), p. 173. Plural; (examples), p. 173. 

IV. Series. External Inflexion; fern, nouns in -e and -i, p. 173. 

Singular; (examples), p. 173. Plural; (examples), p. 173. 
V. Series. External and Internal Inflexion ; fem. nouns, p. 173. 
Singular ; (examples), p. 1 71. Plural ; (examples), p. 1 71. 

Second Order. — The "consonantal", p. 174. 

Paradigm of five series; the first three liquid, the last two mute, p. 

I. Series. Subst. in -im t -in, taking gen. sing, -a or -e, etc. Exam- 
ples, p. 175. 
II. Series. Nouns taking in oblique cases -an, -in, and -in, -en; (two 
divisions), p. 175. Examples, p. 176. 

III. Series. Nouns of relationship, masc. and fem. in -ir, p 176. Ex- 

amples, p. 176. 

IV. Series. Derivatives in -id, declined by variation of internal vowels ; 

(two divisions). Examples, p. 1 77. 
V. Series. Contains fem. nouns in -r, to which are added the suffixes, 
-ach, 'ich, ig, p. 177. Examples, p. 178. 
The Dual Number, p. 177. 
Paradigms of the series of the First Order. Examples from the MSS., 

pp. 178-9. 
Anomalous Substantives. Examples, p. 179. 

(B) Diminutives. Instances from MSS., p. 180. 

(C) Degrees of Comparison, p. 180. 

Comparative. Two Forms. Examples, pp. 180-181. 
Superlative. Two en lings. (Examples), p. 181. 



Indo-European or Primitive 

South Aryan. 

Sanskrit Index . p. 185 

Old Persian ,, . „ 

North- West Aryan. 

Greek Index . . „ 

Italic and Romance. 
Latin Index 
Mediaeval Latin Index 
Picenian „ 

Sabine , , 

Oscan „ 

Umbrian , , 

Romance „ 





Italian Index 



Windic or Liio- Slavonian. 

Proven9al ,, 


Old Slavonic 

p. 195 

French „ 


Polish Index 

. 196 


Servian „ 


Gothic Index . 


Lithuanian Index 

• ;> 

Old Teutonic „ 


Lettish „ 

. 197 

Old High German 


Old Prussian „ 

* j» 

Middle High German Index 



New High German 


Old Celtic Index 


Old Saxon 


Old Irish „ 




Middle Irish „ 

. 204 

Low German 


Modern Irish „ 

• »> 

Middle Dutch 


Welsh „ 




Kyrnric „ 

. 208 



Cornish „ 

• u 

Old Norse 


Armoric „ 





Gaulish and Old 


Kymric Index 

p. 220 



. 213 

Cornish „ 


Irish Index 


Armoric „ 

• » 

Welsh „ . 





§. 1. Of Roots and Root-Forms. 

THE method of investigation employed in the modern science 
of Comparative Etymology may be described as an analytic 
process, to which the words of cognate languages are subjected; 
consisting in successively stripping from them certain letters or 
syllables which have the symbolical power of expressing the quali- 
ties, proportions, or relations in space and time, under which the 
subject contemplates the object — that is, so much of the phonetic 
whole constituting the word, as fixes or limits the idea intended to 
be expressed by it, and makes it the symbol of a definite concep- 
tion. By this stripping process we obtain a residual syllable or 
nucleus to which the term Root (French, Racine; German, 
Wurzel) is given. A large number of different words in the 
same language, subjected to this kind of analysis, may leave the 
same syllable or root ; hence we may consider the Root of a series 
of words as a phonetic symbol of an individual but logically in- 
definite idea, the limitation or logical definition of the idea being 
given by the sounds or syllables stripped off. The assumption 
of such mono-syllabic nuclei in words has given rise to the hypo- 
thesis that the formative process or growth of languages was a 
synthesis, the reverse of our analysis ; or, in other terms, that the 
first symbols of ideas in language were Roots, out of which 
were elaborated the more developed forms and words. 

If we compare the different forms which the same word as- 
sumes in the several dialects of a language, we shall find that the 
difference is due to the substitution of certain letters for others. 
A similar comparative study of all languages, shows us that they 
may be grouped into families, the members of each of which may 
be looked upon as dialects in a wider sense, of some more primi- 
tive language. Although at first sight, the permutations, or letter 
changes from one language to another, appear to be quite arbi- 
trary, they nevertheless take place according to definite laws, 
which are proper to each language. A very good example of 
these phonetic laws is afforded by the remarkable permutation or 
alteration in historical times of the mute consonants in the Teu- 
tonic languages (Lautverscliiebiing), schematized by J. Grimm, 


4 Introduction. 

according to which these consonants appear, in passing from the 
Greek or Latin to the Gothic, and thence to the Old High German, 
to be shifted forward in the direction in which the sounds are na- 
turally developed — that is, the labial, dental, and palatal medials 
pass into the corresponding tenues, and the latter into the aspi- 
rates — thus the Gr. medial b is represented by the Goth, tenuis p 
and by the O. H. G. aspirate pit or/; the Gr. p by Goth. / and the 
O. H.G. 6, etc. ; the Gr. dental medial d by the Goth, tenuis t and 
the O.H.G. aspirate th; the Gr. medial g, by the Goth, tenuis k, 
and the O.H.G. aspirate ch, e.g.: Gr. ttovq, gen. ttoBoc, Goth./o- 
tus, O. H. G. vuoz; Saicpv, Goth, tagr, O H.G. zahar (the sibilant 
z for the aspirate ill) ; Lat. gelidus, Goth. Icalds, O.H.G. chalt y etc. 1 
By the study of the phonetic laws which govern the permuta- 
tions or letter changes in each member of a family of languages, 
we may determine the words in each family which have had a 
common origin. On analysing these words we obtain a series of 
residual syllables, which, like the words from which they were 
obtained, differ from each other, and are nevertheless but forms 
of the same root. The primitive form of the root could only be 
found in the mother tongue of the family ; but as no monument 
of this language has been handed down to us, we can only dis- 
cover this root inductively, by a comparative study of all the lan- 
guages of the family. What we obtain by the analysis of the 
words of a language, are not, therefore, properly speaking, roots, 
but only Hoot Forms. The root forms of the same root may often 
present so great a dissimilarity, that, without a knowledge of the 
permutations of the letters, and a comparison of all the forms in 
a family, we would not suspect any relationship between them. 
Thus the German word wer presents at first sight so little 
resemblance to the Latin one quis, that we could not suppose that 
they were the same word, or even that they contained the same 
root; and yet this becomes evident enough by comparing the 
forms of the word in several languages, which give us the inter- 
mediate links, e. g.: Skr. has; Gr., rig; Lat., quis; Goth., hvas; 
O.H.G., huer; N. H. G., wer. The object of comparative ety- 
mology is to determine first, the root forms, and then the roots ; 
but it also includes that of the grammatical terminations which 
are added to the roots. Comparative Etymology may, conse- 
quently, be considered as a species of Palaeography which has 
for its object the determination, from their mutilated relics, of 

1 I do not profess, in this Introduction, to discuss the value of particular 
laws, my object being merely to explain the nature of Eoots, Stems, etc. I have 
endeavoured to state Grimm's law as simply as possible, but, of course, the form 
in which I have given it is not wholly unobjectionable ; and this the more so, as 
I am aware that some of the examples do not harmonize with Benary's impor- 
tant law. 

On Roots, Stems, and Derivatives. 5 

the primitive forms of a language, — of that of the parent lan- 
guage of a family of languages, — and, ultimately, of the parent 
language of all ; exactly as the object of Palaeontology is to re- 
construct from the bones, shells, etc., the forms which extinct 
animals had when living. 

Leaving out of consideration interjections, we may classify the 
different kinds of words of which speech is composed according 
to the following division, which is that usually followed by gram- 
marians : — 


Noun-substantives. Pronominal substantives (pro- 

nouns, /, tkou, he, she, it, who, 


A. Words defining the subject — Predicate words. 

a. Adjectives. 

«. Qualitative adjectives. b. 1 Quantitative adjectives 

(numerals, etc.) 

2 Pronominal adjectives 

(mine, thine, this, etc.) 

3 Articles. 
/3. Verbs. 

a. Concrete verbs (to love). b. Abstract verbs (to be). 

B. Words defining the Predicate — Adverbs. 

a. Qualitative adverbs b. Adverbs of time, place, 

(derived from adjectives). number, etc. 



This arrangement renders the distinction between the words 
which constitute the materials of speech, and those which express 
the varying relations of space, number, time, etc., very evident. 
And as the words of each class may be subjected to the process 
of analysis, we get two kinds of roots, distinguished also as Cor- 
poral, and Formal or Formational Roots. As we may get the same 
root from a noun, an adjective, a verb, or an adverb, a corporal 
root must be considered to have the embryonic power of a whole 
sentence ; that is, of expressing a whole concrete conception, but 
without possessing any means of expressing the person, time, etc. 
Corporal roots may therefore be considered as germs of nouns and 
verbs, rather than as possessing the explicit power of either. 

All languages may be classified into a few classes, according 
to the manner in which the two kinds of roots are joined to one 


6 Introduction. 

another. We niay, for example, assume three stages of compo- 
sition: 1, Parathesis, or the mere juxtaposition of roots; 2, Ag- 
glutination, or the adhesion of roots ; 3, Amalgamation, or the 
fusion of roots. 

Parathesis. A language at this stage would consist of mono- 
syllabic roots simply, the grammatical relations being expressed 
by juxtaposition with other roots. The same root, according to 
its position in a sentence, may perform the function of a noun, 
an adjective, verb, etc. Pott calls such languages, of which the 
Chinese affords an example, Isolating languages. 

Agglutination. In this stage the grammatical relations — mood, 
tense, person, and class of verbs, number, cases, etc., of nouns, 
are expressed by affixes to monosyllabic roots, which, though 
invariable in function, are not inseparable from the root, the 
several relations being expressed by successively added affixes. 
In some agglutinating languages all the affixes are suffixes : thus, 
in the Finno-Tatarian languages, where the root-vowel, itself 
inflexible, modifies the vowels of the suffixes, giving thereby 
rise to the so-called vowel harmony. Other agglutinating lan- 
guages have apparently almost exclusively prefixes, as the Kaffir 
languages of South Africa. The Semitic languages show a 
higher stage of agglutination by admitting of prefixes as well as 
suffixes, the cases of nouns being expressed by prefixing prepo- 
sitions, 2 and still more by employing internal vowel changes as 
means of inflexion. 3 

Amalgamation. When the corporal and formational elements 
become so intimately blended that both fuse into an indissoluble 
unity, the formational elements give rise to true inflexion, which 
produces a complete logical distinction of the grammatical cate- 
gories. Languages at this stage are called by Pott, Amalgamating. 

Bopp's classification is somewhat different. He makes three 
classes also, the first corresponding to the parathetical ; but in 
the second he includes both agglutinating and amalgamating, 

Do O O O ' 

and makes of the Semitic languages a third distinct class. 

The amalgamating languages are consequently those which 
have the most perfect organization, and include the Indo-Euro- 
pean family of languages, which comprises the Sanskrit, Latin, 
Greek, Celtic, Slavonian, Gothic, and their modern descendants. 
In their primitive state such languages cannot contain uninflected 
roots. In process of time, however, and especially if great per- 

2 The Arabic, however, has real case terminations. 

3 Some examples illustrative of the process of agglutination in the Xorthern 
Family of languages maybe found at pp. 92 and 9±, vol. I., of the Atlantis, in 
the first part of my paper ' ' On the influence which the Physical Geography, 
the Animal and Vegetable Productions, etc., of different regions exert upon the 
Languages, Mythology, and early Literature of Mankind, etc." 

On Roots, Stems, and Derivatives. 7 

turbations and mixtures of different peoples take place, trie gram- 
matical elements affixed to the roots get shortened, mutilated, or 
drop off wholly, so that the root is laid bare. In modern lan- 
guages, as, for example, the English, we find several naked roots, 
which, however, have the value of the words from which they 
have been obtained by the gradual wearing off of the clothing; 
thus the word hand is in reality a root-form, having now the full 
signification of a primitive noun, which in Gothic had the form 

No matter how great the phonetic modifications which a root 
may undergo in producing a number of root forms, it preserves 
its identity. Some philologists, however, admit of exceptions to 
this rule ; that is, they consider that certain phonetic modifications 
of a root may alter its signification so as to produce a new root. 
The process by which this is believed to be effected is called 
Root Variation, and may be described as a phonetic change that 
modifies or tempers more or less the concrete value of the root, 
without the latter ceasing to be a root. The result of this varia- 
tion is to produce in the same language, or in cognate branches 
of the same family of languages, two or more affiliated roots with 
almost synonymous signification, but differing in a slight degree 
phonetically. These synonymous roots may appear to have 
been evolved, as it were, parallel to one another, or the one 
to be primary, and the other secondary. Of two such synony- 
mous roots we may consider the one which has the greatest 
phonetic dimensions to be the secondary root. This hypothesis 
has been so far generalized by some philologists that they believe 
all roots of considerable phonetic dimensions to be secondary 
roots, even where we can no longer detect the primitive root. 
Many, on the other hand, do not admit that such a change can 
at all take place in a root. Assuming, however, that this kind 
of variation takes place, it must do so either: 1, by simple modi- 
fication of one or more letters — vowels or consonants — e.g. y\a<p, 
y\v(j) ; ypa(f), grab; or 2, by the addition of a soimd or sounds — 
e. g. Skr. bhd, to shine, which gives the extended root bhdsh, to 
speak, Gr. 0rj ($r)f±i), to speak, extended roots (pav, to shine, to 
speak (<P<i(f)oq, TricpavaKui), (j>av, to shine ^cpaivoj) ; Skr. root ru, 
to sound, extended root rud, to weep. In the change of the 
root into a stem, to be described further on, there is no such 
modification of the root-idea. 

§.2. Of Elementary Word-formation, and Inflexion. 

Assuming that language was synthetically developed from 
isolated monosyllabic roots, we have next to consider how words 
were formed from roots in the Indo-European, or amalgamating 

8 Introduction, 

languages, to which family the following pages will exclusively 
refer. The development of words from roots may be called 
Word-formation, but the elementary words thus formed must 
undergo further modification, in order to express the varying 
relations of speech. Thus, a Verb must have special contri- 
vances to express time, person, etc. ; and the Noun, number and 
case, etc. This further modification is called Inflexion, or Word- 
bending. The processes by which elementary Word-formation 
and Inflexion are effected are fundamentally the same ; they are — 

1. Internal phonetic change, which can only affect the root- 

vowel, as the change of a consonant would necessarily 
produce a change in the symbolic value of the root. 

2. Addition of phonetic material to the root, which may be of 

two kinds : 

a. Such as springs from the root itself; or Duplication. 

b. Affixes ; which may be Prefixes or Suffixes, but espe- 
cially the latter. These Affixes may be : 

a. Single sounds or syllables, which only are used as 
formational elements of words, having by themselves 
no signification in the fully-formed language, and do 
not consequently occur isolated in it. 
/3. Affixes which possess of themselves a distinct mean- 
ing, and consequently may occur as isolated words 
in the language. 
In the Semitic languages, vowel-change is a predominant 
mode of word-formation and word- inflexion. In the Indo-Euro- 
pean languages it only appears as Ablaut;* that- is, an interchange 
in the body of the root of the primitive pure short vowels, a, i, 
u, but, at a later period, of the newer vowels e and o also, which 
were produced by the softening of the primitive vowels. This 
hind of vocalic change (ablaut) appears to have been a funda- 
mental principle of word-formation in the Teutonic languages. 

The vowel change known as Umlaut is the change or obscura- 
tion of the fundamental root vowels a, o, u, into the impure or 
obscure vowels, a, 6, u, under the regressive assimilating influ- 
ence of i (or u) in the syllable immediately following the root. 
In the Teutonic languages, umlaut by means of u only occurs in 
the Old Norse, in which it has been fully developed ; umlaut does 

4 Wherever special technical terms are invented in any language to express 
certain definite ideas, they should be retained in translating from that language, 
if the laws of euphony of the language into which the translation is made 
at all admit of it. The words ablaut, umlaut, vorlaut, nachlaut, anlaut, inlaut, 
and auslaut are convenient terms, and better than any which could be made out 
of Greek words. I have consequently used them throughout. Ablaut, umlaut, 
vorlaut, and nachlaut are fully explained where they first occur. Anlaut is the 
initial sound, and auslaut is the final sound of a word. 

On Roots, Stems, and Derivatives. 9 

not at all occur in the Gothic. In the Zend umlaut is produced 
by both the vowels (i and u), a becoming ai under the influence 
of an i following, and au under the influence of an u following. 
When the i is softened to e, the umlaut remains as a rule, and 
even is retained when the e is dropt. Umlaut thus apparently 
acquires the flexional signification of the ending, by the action of 
which it was produced, and now acts as its substitute, although 
originally it was a mere phonetic consequence of it. We have a 
good example of this in the preterite of the conjunctive mood in 
the O.H. German; the preterite forms of the strong conjugation, 
which have conditional or potential signification, are charac- 
terized by an i, i. In the M. and N. H. German, this i passes 
into e, but leaves evidence of its existence in the umlaut of the 
root vowel, which now characterizes the conjunctive: O.H.G. 
Pnet. Ind. sing, first, second, and third persons, las, last, las; 
prast. conj. last, lasis, lasi ; M. H. G., laese, laesest, laese. The 
following are additional examples of umlaut: O. H. G., anti 
(enti), M.H.G., Ende; O.H.G., Iwndi, hendi, N.H.G., Hdnde; 
O.H.G., ti'dki, N.H.G., trdge. There is also a phonetic process 
of regressive assimilation, the reverse of umlaut, and which is 
called Breaking, or Fracture, by which i is changed into e, and o 
into u, by the action of an a following. 

The remarkable law of progressive vocal assimilation already 
alluded to, and which constitutes so characteristic a feature of the 
Finno-Tatarian family of 'languages, may be described as a kind 
of progressive umlaut, which it will be useful to describe, as it 
will be alluded to hereafter. In the languages of that family the 
vowels may be divided into three classes : hard, a, o, u, and in 
some languages,^; 2. soft, a, o, u; 3. neutral, i, and in some 
languages, as the Finnish and Samoyede, e also. If the root 
syllable, which is invariable in all the languages of the family, be 
hard, the vowel of the suffixes cannot be soft ; conversely, hard 
vowels cannot follow soft ones. The vowel i, and in Finnish, 
etc., e, also, may be followed either by a hard or soft vowel. 
The Irish rule of " broad to broad, and slender to slender", may 
be looked upon as progressive assimilation; the Irish broad 
vowels being a, o, it, and the slender e, i. Wherever this rule 
is followed, a consonant, or consonants, should in every written 
word lie between either two broad, or two slender vowels ; or, 
in other words, if the vowel of a syllable be broad, the vowel of 
the next succeeding syllable should be broad ; if the vowel be 
slender, the following one must likewise be slender. 

The peculiar weakening of the root vowel which is produced in 
Latin words by the vowel of a prefix, whether due to composition 
or reduplication, may likewise be looked upon as a species of pro- 

10 Introduction. 

gressive assimilation analogous to that which exists in the Irish. 
The Finno-Tatarian languages having no prefixes, all progressive 
assimilation must affect not the root but the endings, hence the 
difference between this phonetic change in the Latin and the 
languages in question. The following examples will show the 
character of the change in the Latin : under the influence of e, 
i, and also o, a becomes i or e, e becomes i, ce passes into I, 
au sometimes into <% into i, u into e — tango, tetigi ; pars, ex- 
pers ; facio, efficio ; placeo, displiceo ; jacio, objicio ; annus, per- 
ennis ; folio, refello ; carpo, decerpo ; cashes, incestus ; ars, iners ; 
lego, diligo; rego, corrigo; quo3ro, inquiro; caedo, cecldi; claudo, in- 
cludo; notusz=.gnotus, cognitus; juro, pejero, etc. There are, how- 
ever, numerous exceptions, and in compound words formed by pre- 
fixed particles or prepositions, such as cirewn, ante, per, etc., it does 
not occur. In ago, abigo, we have a change into i produced by a. 
Phonetic change, by means of affixes, is the great agent in 
word-forming in the Indo-European languages. The first kind 
of affixes are those employed in word-formation properly so 
called, and in inflexion. The second kind of affixes — that is, 
those which possess of themselves a distinct meaning — are used 
in making compound words. Some of the first kind of affixes 
may, however, be distinctly traced to independent words: as 
examples may be mentioned the personal endings of the verbs, 
the signs of many of the cases, etc. Thus the ending of the 
first person in the Sanskrit and Greek was mi: bha-mi, tudd- 
mi, ddsyd-mi; u-jil, (prr/nl, Dor. (fra-pi; in the Latin the i has 
been lost, and the ending is now only m — su-m, inqua-m, dicere- 
m; this mi is the pronominal stem ma softened to mi, as we 
actually find it in mi-Jii. The first person plural ending in the 
Sanskrit is -mas, in the Veda dialect, masi, in the Doric dialect 
of the Greek, jueq: Skr., bhd-mas; Dor., (jya-fiig; in the Latin 
it was mus, and in the O. H. German, mes. From the Veda 
form masi, Curtius considers the ending to be made up of the 
pronominal stems of the first and second pronouns: ma -{-si (si= 
ti); that is, I-\-thou = ive. Again, the Greek endings, -era and 
-cfoj, of the 1 Aorist (typa\pa for e-ypa^-aa), and of the future 
(y(j>a^u) for ypafy-aii)), and of the Latin ending of the perfect, -si 
(scripsi), are obtained from the verb, in the Greek d-jil, Dor. 
IfjLfil, Lithuanian esmi, root as. And lastly, the Latin imperfect, 
-bam, and the future, -bo, are derived from the xootfu (in j iiam, 
fu-turus,fu-i). The English suffixes -ly, -hood, -ship, -some, are 
also good examples, meaning originally like, state (A. Sax. had), 
shape, same. Indeed, the distinction between simple word-for- 
mation and composition cannot be always accurately defined; 
practically, however, it exists in fully formed languages. 

On Roots, Stems, and Derivatives. 11 

If some of the affixes can thus be derived from significant 
words, it is perfectly reasonable that philologists should endea- 
vour to generalize the fact, and assume as probable that all word- 
formino- and flexional affixes, which possess the symbolic signi- 
fication of formational Avords, were originally formed by affixing 
such words to the word to be grammatically modified. In mo- 
dern languages where the flexional endings have been worn off, 
their functions are again performed by words already existing 
in the language. Such a view naturally leads to the assumption 
that in the gradual development of languages all word-formation 
and flexion were synthesis or composition. 

The hypothesis that word-formation and flexion were primi- 
tively synthesis, and that the phonetic additions by which they are 
affected were at first independent words, constitutes the basis of 
what is known as the agglutination theory. This theory is now 
generally considered to be the correct one. Some philologists 
seem disposed, however, to modify it so far as to admit two kinds 
of word- forming and inflexional materials: 1, Simple sounds or 
syllables, which were never words by themselves, their symbolic 
power being derived from that which each individual letter is 
considered inherently to possess; 2, independent words worn out 
into word-forming and flexional elements. 

§. 3. Of Primary Stem-formation. 

In the foregoing sections three kinds of forms have been men- 
tioned: 1, roots; 2, elementary word-forms; and 3, words clothed 
with the inflexional elements, which express their relations to 
each other as members of a sentence. But these do not include 
every form. The simple word-forms are not as a rule obtained 
by the direct addition of a grammatical element, derivational or 
inflexional, to the root. Between the root and the grammatically 
complete word there lies the ivord-stem (French, Radical; German, 
Stamm, and corresponding to the Crude-form of some English 
writers), to which, and not to the root itself, the grammatical 
elements are added. Stem-formation is, consequently, the first 
stage of word-formation, a stem is not a root, nor yet a complete 
word. From the root it is logically distinguished in this, that 
the unlimited, or, as we might say undulating contents of the 
root are fixed or solidified, and rendered fit to serve as a symbol 
of the completely determinate conception represented by the 
grammatical word. While there are but two classes of Roots, 
corporal . and formational, there may be many kinds of Stems : 
for example, we may have verbal, nominal, pronominal, and 
particle Stems — each kind of root branching into many stems, 
according to the grammatical changes it may undergo. Instead, 

12 Introduction. 

then, of three categories of phonetic forms, we have, in reality, 
four : Roots, Stems, being of a two-fold kind, Simple word-forms 
or derivatives, and Words clothed with inflexional elements. 

Word-formation from roots consists, then, of two distinct pro- 
cesses: 1, the formation of stems from roots, or, Stem-formation; 
and 2, the formation of words from stems, or Derivation in its 
simplest form. Both processes are effected by phonetic means to 
be hereafter described, but here it may be useful to mention that 
they cannot always be absolutely distinguished, — the same pho- 
netic change or addition being at one time stem-formation, and at 
another true derivation. There is, however, an essential diffe- 
rence between stems and derivatives, the basis of the true stem 
is the root, while the derivative always proceeds from the stem. 
The two processes are, therefore, logically, even when not pho- 
netically, distinct. 

The Phonetic methods of primary or Pure Stem-formation 
may now be described in detail ; they are : — 
I. Modification of root-vowel. 

1 . Ablaut proper, which is a very frequent change in the Greek ; 

it is rather an accompaniment than a means of stem-forma- 
tion. It does not often occur in the Latin, but in the Teu- 
tonic languages it is very common, and was apparently 
the primitive means of stem-formation. Examples: root 
N.H.G. brack, stems brich, bruch; root ]3aX, stems ]3oA, 
(5eX, verb £-/3aX-ov, nouns /3oX-?7, fi£\-og (tego, toga). 

2. Obscuration of the root-vowels a and i to e, and of u to o. 

The Greek and Latin have no fraction of u, i to o, e, the 
change is always the inverse. As an example of the 
breaking of a to e may be given: root lag, stems Xsy, 
leg, verbs Xiyw, lego; and of i to e, the Teutonic root LIB, 
to remain (zzAnr), Goth, liban, to live = O.H.G. leben. 

3. Strengthening of the root-vowel, which may take place : 

a. By lengthening the short vowel, as : root XaO, stems XrjO, 
XaO, verbs XiiOw, Dor. XaOco, nouns XrjQri, Dor. XaOa. 

b Gunation 5 and Diphthongation — Examples of guna- 
tion: root i, stem «, verb elfiL; root <pvy, stem <pzvy, 
verb (frbvyu) ; in Gothic, root bug, stem baug; root vit, 
stem vait. Examples of Diphthongation: root <pav, 
stem <paiv, verb (paivw ; root rav, weakened root form 
rev, stem retv, verb teivw; root $a, stem <W, verb 
Bald). The latter and similar roots ending in a 
vowel show the true relation of the i to the root- 

5 The term gunation is applied to the process by which e (ai) is produced by 
prefixing a to i or i, or 6 (au) by prefixing a to u or u. Diphthongation and gu- 
nation are well expressed by the German terms nachlaut and vorlaut. 

On Roots, Stems, and Derivatives. 13 

vowel in fyaivw. Curtius. has shown that in the 
latter the form was (pav-i-w, a derivational i (Sanskrit 
ya), being originally placed after the root, but which 
by metathesis afterwards entered the root. Gunation, 
according to some scholars, does not occur in the 
Latin, and consequently the derivational i retains its 
place outside the root in the verbs in io of the third 
conjugation, as capio, morior, etc. This opinion is 
not, however, strictly correct ; for although gunation 
may be rare, the following examples show that it does 
sometimes occur : foedus for foidus (if we may connect 
7r£-TroiO-a.), root fid, wiO, bi(ii)d; anrum, aurora, com- 
pare uro, us-tum, Sanskrit root ush. The occurrence of 
this derivational i as an element of stem-formation 
gives rise to a distinct and important class of stems, 
which will be fully discussed further on in the section 
on ya- or m-stems. 
II. Consonantal strengthening of the root. 

1 . Duplication or doubling of the final consonant. 

In the Greek XX, dialectically pp and vv ; <r<x (Boeot.) per- 
mutated in the new Attic to tt, root airap, cnrEp, stem 
cnrtpp, verb (nreppu 6 ; root Krav, ktw, stem ktevv, verb 
KTtvvii) (Lesbian). In the Latin there is frequent du- 
plication of /, root pal, pel) stem pell, verb pello, and in 
the German of I and m, root seal, O.H.G. scellan. In 
the former case it is the result of the assimilation of a 
derivational y by the final consonant. 

2. Affixation of a mute consonant foreign to the root. In the 

Greek and Latin a r is thus affixed frequently, e.g., root 
|3Xaj3, stem /3Xa7rr, verb ]3Xa7rrw ; root tvtt, stem tvttt, 
verb rvTTTU) ; root pac or pec, stem pect, verb pecto [the 
Greek ktuq however suggests that the ct of pecto may be 
radical]. In the Teutonic languages this process is not 
now recognizable. 

3. Affixation or intercalation -6f a nasal. 

a. Nasalizing an internal vowel. This change is common 

in the Latin — e.g., root pag, stem pang, verb pango; 
root liq = Xnr, stem li?iq, verb linquo; root frag, stem 
frang, verb frango; it also occurs in the Gothic, root 
stap, inf. standan, Engl, stand, stood. 

b. Affixation of the nasal in the auslaut: 

a. After vowels. In the Greek we get from ra, ya, 
rav, tsv, yev. It is sometimes combined with diph- 

6 Such would have been the Lesbian form, at least ; Attic, aireipo). 

14 Introduction. 

thongation, as in root /3a, stem fiaiv, verb fiaiva), 

In the Gothic we have ga becoming gang. 
]3. After consonants. Only few examples in the Greek ; 

e.g., root rap., tb/ul, stem tbjliv, verb tb/ulvu). In the 

Latin we have sterno, sperno, etc. 
c. Affixation of a whole syllable, accompanied by nasali- 
zation, of which we can only find examples in the 
Greek, e. g., vb, va, root Sa/u, stem Sa/uva, verb ^afi- 
vaw ; as av, by which the root- vowel becomes likewise 
nasalized, root \a6, stem \av9av, verb Xavdavu). 
4. Reduplication . root, jiva, reduplicated stem, incho- 
ative verb fjLi-jiv{]-GK-(jL). Lat, root min, men = Sk.T. 
man=inva? stem me-min, verb memini. 

All the more important methods of primary or pure stem-for- 
mation are embraced under the preceding categories. There 
are also a few exceptional cases, such as - where an intensive s is 
introduced into the root, e. g., root fiiy, stem /uiay, verb fiicryto; 
Latin, misceo (compare Ir. cummasc, commixtio), which must not 
be confounded with the derivational sc of inchoative verbs. Be- 
sides primary or pure steins, there are, however, other classes of 
stems which are formed by the addition of a vowel, or of a 
syllable ending with a consonant, and which will be described 
hereafter. It may be well to observe here, that the circum- 
stance of stems being formed by the addition of a whole syllable, 
the introduction of an intensive s into the body of the root, etc., 
shows us how cautious we should be in concluding that stem- 
forms, which at first sight appear extremely simple, are the roots 
themselves. For example, (pav and Kpiv, although apparently 
forms of very moderate phonetic dimensions, have been, in 
reality, enlarged from <pa and Kpi. Then again, it is necessary 
to be careful to distinguish between the stem and the pure words 
or stem-words. For example, \irog and corpus are true stems, 
as is shown by attaching flexional elements to them ; thus, i irsa- 
og contracted to stteoq, corpus-is softened to corpor-is. On the 
other hand, (pL\i(w), Xoyo(c), fructu(s), are full words, contain- 
ing the derivacional elements, w, c, s, respectively. 

§. 4. Of Noun- formation, and Derivative Steins. 

The formation of stems may be considered the first separation 
of words into grammatical categories, but it does not complete it ; 
for although some stems are essentially verbal, and others nomi- 
nal, there are many which admit of being made the basis either 
of verbs or of nouns. The complete separation is only effected 
when one of the signs which characterize the complete ward is 

On Boots, Steins, and, Derivatives. 15 

affixed to the stem. These grammatical signs are the derivational 
and inflexional elements. 

The characteristic signs by which the stem becomes a verb, 
are the personal endings fit, m, ri ; m, s, t, etc. : root i , stem 
u, verb ei/ul, — jil being the personal ending for the first person 
sing. ind. ; Skr. root tud, stem tuda, verb tuda-si, — si being 
the ending of the second person sing. ind. As Dr. Ebel's 
paper does not deal with the verb, I shall confine myself exclu- 
sively to nouns. The characteristic signs by which nouns are 
formed are the gender and case endings. The vocative, from 
its nature, ought to present us with the pure nominal stem, but 
in the actual language this is not generally the case ; and hence 
it is found more convenient to assume the nominative as the basis 
of analysis. 

One of the most characteristic distinctions between objects is 
that which life affords, and accordingly the sign, by the affixation 
of which to the stem the nominative form of the noun is pro- 
duced, is a gender sign. For living objects, the sign primitively 
affixed to noun-stems in the Indo-European languages was s. 
Some scholars hold that neuter nouns were distinguished by t t 
which they consider possesses a certain power of symbolizing 
lifeless or inert bodies. But the evidence that t was ever used, 
except in pronominal declension, as a sign of the neuter gender, is 
very doubtful. The Gothic neut. adjective-ending ata is, accord- 
ing to Bopp, merely a suffixed pronoun. Mankind has, however, 
at all times, figuratively endowed certain lifeless objects with life, 
and abstract conceptions, such as justice, virtue, etc., are ex- 
pressed by words of masculine or feminine gender, according as 
our fancy chooses to consider them of the one or other sex ; the 
names which are used to symbolize these objects or abstract con- 
ceptions take, accordingly, the sign of living objects. 

The nominative sign s has, however, been but imperfectly pre- 
served ; in the Sanskrit it is usually softened to h ; the feminine 
forms, which incline to vocalic auslaut with long vowels, seem 
to have thrown it off, apparently with the object of marking 
the distinction of the sexes. This tendency to have vocalic 
auslaut is well shown in the adjectives having the endings in 
the Sanskrit, as, a, am; in the Greek, oq, a (17), ov. Even mas- 
culine forms often lose the s. In the Zend and Lat. it is frequently 
dropped altogether. In the Gothic it is generally only preserved 
in masculine substantives with vocalic stems, and in masculine 
adjectives and pronouns. In the O.H.G. the substantives have 
altogether lost it,while in adjectives and pronouns it has become r. 

The neutral t of the pronominal forms has to a great extent 
been lost. In the Greek it does not occur at all : in the Latin it has 

16 Introduction. 

become d : id, Mud, quid, etc. In the Gothic it occurs in the 
pronouns is, si, ita; English he, she, it; Old Irish e, si, ed; Gothic 
sa, so, thata, Anglo-Saxon se, seo, thdt = Greek 6, q, to for ror= 
Sanskrit tat. In the O. H. German it becomes z: Gothic third 
person of the pronoun masc. is, neuter ita = 0. H. German masc. 
ir sometimes er, neuter iz, sometimes ez. In the Gothic blindata, 
godata, O.H.German plintaz, guotaz, M.H. German blindez, guo~ 
tez, the ending ata, as above observed, is a suffixed pronoun, 
and cannot consequently be considered as a proof that t was the 
sign of the neuter, in other than the pronominal declension. In 
many cases the neutral t has been replaced by m or n, which, 
however, belonged originally to the accusative singular. 

The grammatical signs or endings cannot always be directly 
affixed to stems ; this is especially the case with those beginning 
with a consonant, and where the stem ends consonantally. If 
in such cases the ending were affixed directly, the final stem- 
consonant would be rendered liable to change, and the mo- 
dification may proceed so far as to render the stem unrecogni- 
zable. Therefore a vowel is introduced between the stem and 
the ending, which originally had a mere phonetic function, 
and possessed no etymological or grammatical signification. 
The vowel by itself is always short, and consesequently very 
changeable. It is often an extremely difficult problem to dis- 
tinguish between the vowel thus added and a derivational 
vowel, and therefore between a derivational and stem-form. It 
is also an important one : for this vowel, though originally having 
no stem-forming or derivational character, has gradually come 
to be looked upon as an integral part of the stem - ending, 
and has even penetrated where it was not absolutely required. 7 
It will be useful to call this stem-forming vowel in nominal 
stems the Declension Vowel, in order to distinguish it from a 
second vowel which is sometimes used as a mere copulative in 
the oblique cases, and which is never an integral part of the stem- 
ending. A similar stem-forming vowel is found in verbal stems. 

Forms which must be looked upon as true stems are, however, 

7 In Finnish nearly all the stems are two-syllabled. The first or root syllable 
is accentuated, the second has a short vowel auslaut. This short rowel, unlike 
the root-rowel, which is invariable, sounds differently according as the stem is 
pronominal or verbal. It is a mere rhythmical addition to the root which some- 
times acquires the signification of a derivational suffix, and has consequently a 
striking analogy to the declension vowel of the Indo-European languages, and 
makes Finnish stems appear very much like those in the Gothic, which will be 
described further on as vocalic middle forms. The affixation of this vowel is 
the only mode of stem-formation in the Finnish; in Hungarian it has been to a 
great extent obliterated. It would be extremely interesting to trace this rhyth- 
mical stem-forming vowel through the whole Finno-Tatarian Family. Here 
however, it would be out of place to dwell further on the analogy. 

On Roots, Stems, and Derivatives. 17 

sometimes made by the addition of a whole syllable, the consonant 
forming the auslaut, taking the place of the declension vowel. 
The forms produced in this way approach nearer to the character 
of derivatives than those obtained by the mere addition of the 
declension vowel, — indeed many of them have the character of 
true derivatives. 

In the Greek nouns in ttjc, we have a perfectly analogous class 
of stems formed by the addition of a syllable ending vocally 
instead of consonantally ; they are, however, in part undoubted 
derivatives formed by affixing the derivational suffix rn to an 
already fully formed stem. We may call all such stems, formed by 
the addition of a syllable to verbal or nominal stems, which thus 
perform, as it were, the function of roots, Derivative Stems, and 
treat of them as a distinct class of vocalic or consonantal stems, 
according as the suffix ends in a vowel or a consonant. But 
as there is a real logical distinction between the true stems 
which start from the root, and these pseudo-stems which are 
derivatives of true stems, it will be better not to consider such 
pseudo-stems under the head of stems, but to refer all of them 
to the category of derivatives. 

We have accordingly three distinct classes of true nominal 
stems as regards their relations to the grammatical endings: — 
1, primary or pure stems, to which the nominative s is directly 
affixed; 2, stems which require a vowel between them and the 
ending ; and 3, stems formed by the addition of a syllable end- 
ing in a consonant in place of a vowel. The second and third 
classes may be called secondaiy stems or, better still, middle 
forms, that is, intermediate between pure stems and true deri- 
vatives. Of the pure stems some have vocalic and some conso- 
nantal auslaut. The middle forms, produced by affixing a de- 
clension vowel, may all be looked upon as vocalic ending stems, 
while the middle forms, which result from affixing a consonantal 
ending sjdlable, are consonantal stems. 

The following table, to which I have likewise added the 
derivative stems, will render the classification of stems above 
given more intelligible : — 

I. True Stems formed from the Root. 

Vocalic Stems. Consonantal Steins. 

1. Pure Stems. 1. Pure Stems. 

2. Middle Forms produced by: 2. Middle forms or stems produced 

a. affixing the declension vowels, by affixing a syllable ending 
a i, u. consonantally to the root. 

b. ya-stems, or, a-stems, with an 
intercalated i (y) before the de- 
clension vowel. 

18 Introduction. 

II. Pseudo, or Derivative Stems. 

3. Stems formed by the addition of 3. Stems formed by the addition of 

a derivational suffix ending in a a derivational suffix, ending 

vowel, to an already fully-formed consonantally, to an already 

stem. fully-formed stem. 

§. 5. Of Vocalic Stems. 

Pure Stems. All monosyllabic nouns may, strictly speaking, 
be considered to be pure stem-words, in which the nominal sign 
is directly affixed to the stem without any intervening phonetic 
material. Such nouns occur in the Greek and Latin, though 
they are not numerous. 

greek: root /a, stem k7, noun kl-q (masc. gen. ki-6q), ypavg 
(ypa-oc) ; SpvQ, Qd)Q (roots Spy, Oo, stems Spy, #a>) possess still 
more of the character of pure stems. Some forms usually included 
under this category are undoubtedly not primitive pure vocalic 
steins ; for example, (3ovg may perhaps be more properly reckoned 
among the consonantal stems, as it stands for )3of-c (root bo). 

latin. In the Latin there are extremely few forms which can 
be considered, strictly speaking, as pure vocalic stems. Perhaps 
the only form is grus, stem gru, for it is doubtful whether the r 
in the plural vi-r-es of vis (stem vir?), — and in the old form of the 
genitive sueris (su-er-is) o£sus, Sanskrit, su-kara, — be not organic 
instead of being, as is generally supposed, merely euphonic. 

ootric. In the Gothic a number of such monosyllabic 
words, belonging to what is called the strong declension, is to 
be found ; in the masculine and feminine they have the nomina- 
tive sign s, while in the neuter no suffix can be found, and the 
stem accordingly occurs in its naked form, e.g. : masc. jisk-s, 
dag-s, balg-s; fern, anst-s; and neut. leik. These nouns corres- 
pond with the Greek nouns derived from consonantal stems: 
6pi%, a'/£, 7rvp, and the Latin nouns urb-s, pon-s, met. In the 
nominative case, the analogy is complete; but if we compare 
them through all their cases, we shall find that in the Greek 
and Latin the nouns of this kind affix the case-endings to the 
stem in exactly the same way throughout, namely, its nomi- 
native directly, and the others by means of a copulative vowel, 
which is the same in all the cases, while the Gothic nouns 
take different vowels in the plural. For example : 

Nom. and Voc. . . fisko-s balge-is 

Gen. . . fiske balge 

Dat. . . fiska-m balgi-m 

Ace. . . fiska-ns balgi-ns. 

It would appear from this, that the Gothic nouns under con- 
sideration are only relics of more primitive forms, still preserved 

On Boots, Stems, and Derivatives. 19 

in the plural, but blotted out in the singular. According to this 
view, all the nominal steins must have been clothed with a voca- 
lic auslaut, which was either a or i, and called by Grimm the 
Declension Vowels, a term which I have extended above to the 
corresponding vowels of the vocalic middle forms in the Greek 
and Latin. The primitive form of fisk-s must therefore have 
been fiska-s, and of balg-s, balgi-s — forms which approach very 
close to the Latin, as may be seen by comparing the primitive 
form of gast-s, gasti-s = Latin, hosti-s. The view just put for- 
ward is supported by the circumstance that there exists a class 
of nouns, in which the clothing or declension vowel of the stem 
is u, that are not syncopated like those with the vowels a and i. 
Although at first sight the Teutonic languages appear to contain 
the largest number of pure stems, the preceding considerations 
apparently show that there are no pure nominal stems in those 
lansruaffes. On this account I will include the whole of those 
Gothic nouns under the middle forms with vocalic auslaut. 

Middle Forms ending vocally. The term middle form 
implies that we have passed beyond the stem, but have not 
yet arrived at a true derivative. The nouns derived from 
those middle forms have the same analosrv to those obtained 
from pure stems, that the Greek verbs m au), cw, ioj, etc. — as 
TLfiad), <pi\(w, fieOvii), — have to some of those in jull, — as elfd, 
<pr)jj.i, etc. The nominal middle forms have, however, much 
less of a derivational character than the verbs above named; 
so that, while always bearing their mode of genesis in mind, 
we may consider them as vocalic stems. 

As the primitive vowels were a, i, u, — 'e and o having been 
formed later, — the primitive stem-forming vowels must have been 
also a, i, u. To these were added at later language-periods e and 
o, — e being formed by the softening of a, e of i, and o of a. There 
is also a secondary u produced from a, which must not be con- 
founded with the primitive u. We may consequently include 
all vocalic steins under a-stems, z-stems, and w-stems. 

i-stems. — I-stems approach closest to the character of pure 
stems. In the Latin the i often becomes e; in the Greek it 
sometimes becomes in the oblique cases e. 

greek. — The masculines and feminines of the third declen- 
sion in -ig, gen. -tog, -£ioc, belong to the z-stems ; e.g., (pvaig (-tog) 
ttoXiq. There are no neuter nouns in i. The adjectives like 
'ISpig, ISpi are of this class. 

latin. — The masculines and feminines of the third declension 
in is, and the feminines in es, which do not take an augmenting 
syllable in the genitive, belong to the z-stems, as, liosti-s, civi-s, 
aede-s; the i being changed in the latter into the long e charac- 


20 Introduction. 

teristic of feminines. In neuters the i is changed into e, hut in 
the plural the i again appears: mar-e, mar-is, mar-i-a. The 
stems, brevi, dulci, levi, of the adjectives, brevis, dulcis, levis, 
belong to this category. 

gotsic. — Among the Gothic z-stems which correspond to the, 
preceding, may be mentioned the primitive forms : masc, gastis, 
gardis, balgis, existing in the Gothic, in the syncopated forms : 
gasts, gards, balgs, but showing traces of the vocalic clothing of 
the stem in the plural: nom., gasteis, gardeis, balgeis; fern., 
dedis, vaurtis, syncopated in the Gothic to deds, vaurts; nom. 
plur., dedeis, vaurteis. It will be seen from the preceding, that 
the feminines also retain the nominative sign s, the feminine form 
appearing to be marked by a gunation of the vowels of the 
endings in the genitive and dative singular, thus : 

Masc. Fern. 

Nom. . . gast-s ded-s 

Gen. . . gast-is ded-ais 

Dat. . . gast-a ded-ai 

As in the Greek there are no neuters formed from {-stems. 

In addition to the feminines above discussed, and all of which 
belong to the strong declension, there is another peculiar class of 
{-stems belonging to feminine nouns of the weak declension, such 
as 7nanagei, gen. manageins, which will be better understood 
when I treat of the a-stems. 

Adjectives derived directly from stems, and not through other 
forms, although differing essentially from substantives in their 
flexion, exhibited primitively the same distinction of stems into 
a-, i-, and ?/-stems, corresponding to the Greek adjectives in og, 
a, ov, and vg, eta, v ; and to the Latin in us, a, w, and in is, 
e. But the primitive distinction is very much obscured in the 
Gothic, in which, with the exception of traces, the {-stems have 
wholly died out, while only a few of the ^i-stems remain ; and 
even these pass in the oblique cases into the a-stems, with the 
addition of a derivational i (see the discussion of this subject 
under the head a-stems), so that the primitive character of the 
stem is only recognizable in the nominative. In the Gothic 
adjective there are consequently only a- and w-stems to be dis- 

a-stems. — greek and latin. — The a-stems in the Greek and 
Latin admit of being divided into two classes : — 

1. Stems in which the primitive a has been preserved un- 
changed, or changed into e, and which maybe subdivided into : 
a. Stems with primitive short a. 
/3. Stems with a or e. 

On Roots, Stems, and Derivatives. 21 

2. Stems in which the primitive a has been changed into o in 
the Greek, and into u in the Latin. 

A-stems with primitive short a. In the Greek the masculines 
of the first declension in -ag, -vg are referred to this class, e.g., /3o- 
pzag, 'Epfieiag contracted to 'EpfiriQ ; the nominative g is retained, 
but the vowel is inorganically lengthened. In the Latin, also, onl y 
the masculines of the first declension, which, like the feminines of 
the same declension, have lost the nominative s, belong to this ca- 
tegory, as, scriba, agri-cola, etc. Pott considers the long vowel as 
the result of contraction. It is probable that all the words belong- 
ing to this form are, in reality, derivatives in the second degree 
from nominal and verbal stems, that is, they contain, besides the 
nominative s, a second derivational element, which may still be 
recognized in the Greek nouns in tt]q, as, 7ro\iTr}g, etc., in which 
the derivational suffix is the syllable ttj, as has been already 
pointed out in discussing the different classes of stems at p. 17. 

A-stems with ci or e. As was stated above, feminine nouns 
prefer long vowels and vocalic auslaut ; accordingly we find that 
this class includes the feminines of the first declension in the 
Greek and Latin, all of which have no nominative s. In the Latin 
the a is invariably shortened, but in the Greek it is partly re- 
tained, or changed into rj and partly into a, e. g., x^P"? Sikij, 
<T<pvpa. While the vocative of the a-stems, with primitive short 
a, appears as a rule with the organic short a, that of the stems 
with a or e is the same as the nominative, and consequently 
sometimes has an inorganic short a whenever the nominative has 
one. The primitive long vowel has been preserved in the form 
e, and likewise the nominative s, even in the vocative, in the 
nouns from stems of this class, which belong to the Latin fifth 
declension, which is but an older form of the first, e.g., di-e-s, 
fid-e-s. Here also we meet with forms which appear to belong 
to the class of vocalic stems obtained by means of a derivational 
syllable-suffix, as described above, the analogy being strongly sup- 
ported by their admitting of being declined either according to the 
first or fifth declension, e.g., materies or materia, canities or canitia. 

A-stems, in which the primitive a has been changed in the 
Greek into o, and in the Latin into u. This change occurs in 
the words of the second declension in og, ov, and its, um ; those 
in og and us are, as a rule, masculine (as in the Sanskrit is always 
masculine) ; there are some, however, exceptionally feminine, as 
17 TCKpoog, fagus, etc. The vocative shortens o, u, to e, has orga- 
nically no nominative s, and in the neuter is the same as in the 
nominative. To this category belong the Greek adjectives in 
oe, a (r?), ov, and the Latin ones in us, a, um. From this it 
will be seen that the vowel is shortened in the feminine in the 


22 Introduction. 

Latin, but not in the Greek ; but, on the other hand, some Greek 
adjectives of this category do not distinguish the feminine at all. 

gothic. — To the Gothic a-stems belong the masculine, fe- 
minine, and neuter forms corresponding to the Greek forms in 
oc, a, ov, and the Latin ones in us, a, um, discussed above, and to 
the Sanskrit in as, a, am. For example : masc. dags, fisks, etc., 
which are syncopated forms from dagas, fiskas, etc., as I have 
already fully described, nom. plur. dagos, fiskos, etc. ; fern, giba, 
bida, etc., nom. plur. gibos, bidos, etc. ; neuter, vaurd, leik, etc. ; 
nom. plur. vaurda, etc. The masculines have lost the a in the 
singular, but retained it in the form of o in the nominative 
plural, e in the genitive, and a in the dative and accusative (see 
declension ofjisks, pp. 18, 25) ; the feminines have retained the a 
in the oblique cases as a or o, but have no nominative s. The 
neuter form has lost the a in the singular as well as the neuter 
nominative sign, if it ever had such : — the full form of the nomi- 
native singular of vaurd, for instance, should have been vaurdat, 
if we admit t to be the neutral sign of nouns, more probably 
it was vaurdam = Latin verbum; it has retained it in the plural 

I have already spoken of the Gothic adjectives, and here it is 
only necessary to add that, although the distinction between the 
clothing vowels of the stems was earlier obscured, and to a greater 
extent in the case of the adjectives than in that of the substantives, 
the signs of the genders have been much better preserved. In- 
deed, in the latter respect the Gothic adjectives belonging to the 
a-stems have endings of a much more primitive form than either the 
Greek or the Latin, or even than the Sanskrit. 8 These endings 
are s, a, ata, or, in the primitive form, s, a, t, as for example : — 

Masc. Fern. Neut. 

Primitive organic form . . bhnd-a-s blind-a blind-a-t 

Syncopated Gothic form . . blind -s blind-a blind- a-ta 

ya- (ya-) or ia-stems. — Besides the primitive a-stems, above 
described, there exists another class of stems, which, as they do 
not give rise to any essentially peculiar flexion, may be con- 
sidered as a class of secondary forms of the simple a-stems. 
They are formed by the intercalation of an i (y) between the 
stem and the declension- vowel, and may accordingly be dis- 
tinguished" as ya- {yd-) stems. 9 In the Gothic the stems of this 
class are usually considered to be middle-forms, properly so 
called, the Gothic a-stems being reckoned as pure sterns. I 

8 On the assumption that t was the primitive neuter gender sign, which is not, 
however, generally admitted. Bopp considers the ending -ata to be a suffixed 
pronoun. See § 4, p. 15. 

9 Ya- masculine and neuter ; Yd- feminine. 

On Roots. Stems, and Derivatives. 


think I was justified, however, in classing them along with the 
middle forms of the Greek and the Latin, and, this being so, in 
considering that the ya- {yd-) stems approached still nearer to 
true derivational forms than any of those yet mentioned. 

The Greek and Latin forms which come under this category, 
are the substantives and adjectives in tog, la, iov, ius, ia, 
ium — e. g., Kvpiog, 'idiog ; filius, medius, media, medium. 

As the only forms of this kind referred to in Dr. Ebel's paper 
are Gothic, and as the object of this sketch is merely intended 
to elucidate that paper, I will not further consider the Greek 
and Latin ia-(id-)stems \i)a-{y&-) stems], and will accordingly 
confine myself to a few observations upon the Gothic ones. 

In the Gothic the intercalated y is firmly retained before the 
endings through all the cases, e. g. : haryis instead of hari-s 
(the primitive organic form of which would be haria-s) ; gen. 
haryis; dat. harya, etc. In the Old High German the nom. sing, 
alone retains it. If the stem-syllable be long, or ends in two 
consonants, yi changes into ei: Goth, hairdeis instead of hairdi-s 
(the primitive form of which would be hairdia-s) ; O. H. G. hirti, 
gen. hirtes; M. H. G. hirte, gen. hirtes. In the Modern High 
German it passes into the weak declension: der hirte, des hirten. 
Some of the feminines of this category retain the full organic 
form, such as vrakya, braky a; and in some the i is even preserved 
in the Old High German as y or e, as for example, suntya; 
while other words throw off the a, e. g., Goth. — bandi, kunthi, 
instead of bandya, kunthya. 

The distinction between the a and the ?/a-stems disappears in 
the masculines and feminines when the a and i of the Old Higrh 


German are softened into e, — the two forms then coinciding; 
the existence of such a distinction being only betrayed by the 
umlaut of the stem-vowel in the ya series. The same observation 
applies to the neuters, one example of which will suffice to show 
their forms, e.g.: Goth, kuni, gen. kunyis, instead of the full or- 
ganic form kunya-t, or kunya-m, or kuny-a; O.H.G. chunn-i, gen. 
chunn-es, the i being dropped, as was already noticed in the case 
of the masculines, in all the cases except the nom. sing. In the 
Middle High German the i becomes & as in the other genders : 
kunne, in which the umlaut of the stem-vowel betrays the pa- 
stern. In N.H.G. frequently even the final e is dropped, as, 
Goth, badi (from *badya), O.H.G. betti, M.H.G. bette, N.H.G. 
bett, Engl. bed. 

The adjective forms of the ya-stems are exactly analogous to 
the substantives. In the Middle and New High German the cha- 
racter of the stem is betrayed only by the umlaut. The striking 
analogy between some of the Gothic and Latin adjective forms 

24 Introduction. 

of the ?/a-stems, is well shown by the following comparison: 
Lat. — medius, media, medium = Goth. — midis, midya, midyata. 


stems. — Some Latin ?z-stems drop the n in the nom. sing., e.g. : 
in on, — homo, ordo, margo, of which the full forms with the no- 
minative s should be, homons, ordons, margons; in on, — leo, 
latro, carbo, the full forms of which should be, leon-s, latron-s, 
carbons; the feminine verbal nouns in io, which is obviously 
ion, with the n dropped — actio, ratio, statio, etc., the full forms 
of which should be, actions, rations, stations. These nouns 
give us in the inorganic form of their nom. sing, apparent 
vocalic stems. A comparison between the full organic forms of 
the verbal nouns, which are undoubted derivatives in the second 
degree, and those of the other examples given above, affords 
strong grounds for believing that the latter also are derivational 
forms of the second degree. Except in not having a final n in 
the nom. sing., these nouns are perfectly analogous, in all the other 
cases, to the Latin nouns in in, especially to those in which the i 
is softened to e in the nom. sing., e. g., pecten, etc. ; and the verbal 
nouns flumen, tegmen, lumen, carmen, etc., and may be compared 
with the Greek cucrig, gen. clktXvoq ; \ijiy\v, gen. Xifiivog ; arjdtjv, 
gen. arjSovog; zlkwv, gen. zlicovog; \elju(i)v, gen. Xei/uwvog. 

There is a class of Greek nouns, chiefly feminine, which 
at first sight appear to form their stems in w, and which, as a 
rule, do not take the nominative s, e.g.: 17 ttuOco, gen. 7rei06og, 
the 0) being shortened to o ; 17 ?/x&>, S en - ^X°°£' e ^ c - Some are, 
however, formed with the g , as 77 alSwg, gen. alSoog, the w being 
shortened; ?jpwc gen. ijpbyog, etc., without the shortening of the 
ii). According to Curtius, all these forms are the relics of muti- 
lated w-stems. 10 There is an obvious difference, however, be- 
tween them and the Latin forms homo, etc., with which, if this 
hypothesis be correct, they would connect themselves, namely, 
that the n appears regularly in the oblique cases of all the Latin 
nouns, not only of those ending vocally in the nominative, but 
even of those which take the nominative s, as sanguis, which is 
evidently for sanguins. 

In the Gothic, a class of nouns with vocalic auslaut is also found, 
which exhibit a remarkable analogy with the Latin nouns just 
discussed; for example, guma, Eng. g(r)oom, gen. gumins, which 
may be equated with the Latin homo, gen. hominis; rathyo, gen. 
rathyons, with the Latin ratio, gen. rationis; namo, gen. namins; 

10 This hypothesis of Curtius, by which w, wc, ag, ar, are considered to be=: 
a v, is, to say the least, extremely improbable. Ahrens is more likely right in 
rsgarding aidug , etc., as original c-stems, to which a y (i) is superadded. 

On Roots, Stems, and Derivatives. 


nom. plur. namna, with the Latin nomen, gen. nominis; nom. 
plur. nomina. The reasonable conclusion from this is, that these 
vocalic forms are in reality consonantal n-stems, having more or 
less of a true derivational character. According to this hypo- 
thesis, their full nominative forms should be, guman-s, rathyon-s. 
This hypothesis receives considerable support from the fact that 
several of those forms have again taken up n in the Modern High 
German, e. g. : 



Old High 

bogo, . . 
grabo, krapo, 


Middle High 

grabe, . 


Modern High 

. bogen. 

. graben. 

. garten. 11 
(name and 
^also namen. 

Probably all the foregoing examples may be referred to 
n-stems ; but there is likewise a class of feminine nouns, which, 
considering them as vocalic stems, may be classed as z-stems, 
and which in the Gothic end in the diphthong ei, e.g.: audagei, 
managei, gen. manageins, etc. ; they present the same peculia- 
rities of inflexion as the others above mentioned, as will be shown 
further on. In this case also we are led to the conclusion that 
they are w-stems which have thrown off the n, not only by the 
analogy of inflexion, but also by the fact that the greater part of 
this class of nouns take up an n in the nominative in the O. H. 
German; we thus get, along with maniki, manakin, while in 
the N. H. German we have menge, unlike the a-stems. So also 
O.H.G. odhin and odi, N.H.G. oede, O.H.G. sterchin, N.H.G. 

The dropping of the n does not, as has been already remarked, 
affect the declension of the Latin or Greek nouns ; but it is not 
so in the Germanic languages, where a peculiar declension has 
been developed, known as the weak declension, in contradistinc- 
tion to the strong or true declension of words like fisks, dags, etc. 
The difference will be better understood by the following com- 
parison : 

Strong : Nom. sing, fisks ; gen. fiskis ; dat. fiska ; ace. fisk ; plur. nom. fiskos. 
Weak: „ nana; „ hanins; ,, hanin;,, hanan; „ hanans. 

All the nouns of the class we have been here considering 

11 Besides garda, there is also in the Gothic the word gards (plural gardeis)= 
house, family, etc. ; but evidently having the meaning of garden also, as is 
proved by veingards=vineya,rd ; aurtigards=orch.Sird. The German garten= 
English garden, could not, however, be obtained from it ; but, on the other hand, 
the English yard (as in court-yard) is derived from it. 

2Q Introduction. 

belong to the weak declension, the great peculiarity of which is 
the addition of an n to all the endings of the cases, except the 
nominative singular and dative plural. It belongs to adjectives 
as well as to substantives, but while the latter decline exclusively 
strong or weak, adjectives may be declined according to either 
declension. The weak adjective declension corresponds with that 
of the substantive ; its chief peculiarity is that of having in the 
nominative singular vocalic auslaut in all three genders, e. g. : 

p .* . (Masc. fern, neut. 

(blinda, blindd, blindo. 

The same vowels characterise the genders of the substantive^ 
e.g.: masc. hana; fern, tuggo; neut. hairto. In the Old High 
German the masculine a and the feminine change to and a. 
In Middle and New High German both the a and become e, so 
that all genders end alike. This change is not, however, con- 
fined to the vowels ; for although in the Gothic the case-endings 
are not affected by the addition of the n, the genitive s is dropped 
in Old High German, and ha?iins becomes hanin. In the Middle 
High German, the uniform ending en took the place of all the 
various endings, both singular and plural, with the exception of 
the nominative singular. 

The existence of the s in such Latin forms as sanguis (for 
sa?iguin-s), which belong to the same class as ratio, no-men, 
etc., justify, as I think, the additions of that nominative sign, in 
reconstructing the full organic nominative forms of those and 
similar nouns. For its addition in the analogous German nouns, 
I have the great authority of J. Grimm ; but Bopp's discovery 
that the primitive nominative sign in the Indo-European 
language was s, places the matter beyond doubt. It is right, 
however, to state that some philologists, amongst others Heyse, 
consider that the full organic forms never had s. A full discus- 
sion of this point, however important, is incompatible with the 
limits of our space, and would be in other respects foreign to the 
specific objects for which this introduction has been written. 

u-stems. — greek. Under this head come the Greek words in 
vg of the third declension, which retain the v in the oblique cases, 
e.g.: nom. 6 \\Bv-g, voc. \x®v, gen. \ydv-og, etc., neut. clottv. 

latijv. The Latin w-stems belong exclusively to the words 
declined according to the fourth declension, such as those in its: 
they are chiefly masculine, but also exceptionally feminine, e. g., 
manus, socrus, etc. ; verbal nouns in tics, which may be con- 
sidered to be true derivatives in the second stage, and to which 
the observations made at p. 17 respecting derivative stems con- 
sequently apply, e.g., ductus; neuters in ft, e.g., covnu. The nouns 

On Roots , Stems, and Derivatives. 27 

of tlie second declension, which appear to contain i^-stems, are 
a-stems, the a having been replaced by u. This secondary u is 
much more unstable than the primitive u of the fourth declen- 
sion, which is never suppressed by the vowel of the ending, but, 
on the contrary, absorbs the latter in the genitive singular and 
nominative and accusative plural, e.g., fructus, instead of fruc- 
tuis, fructues. It has not wholly resisted modification, however, 
having been, in most cases, softened into i in the dative and ab- 
lative plural, e.g., from the older fructub us, has come fructibus; 
in others, however, it has remained unchanged, as in acubus, 
lacubus. The whole declension may be considered as a con- 
tracted secondary form of the third declension. 

gothic. The Gothic words founded on w-stems correspond 
exactly with the Greek words in vg of the third declension, and 
the Latin ones in us and u of the fourth. Unlike the Gothic a- 
and z-stems, the w-stems are not syncopated, and consequently we 
get them in their primitive organic forms, the masculine and femi- 
nine taking the s in the nominative singular, e.g.: masc. vulthus, 
sunns, nom. plur. sunyus; fern, liandus, nom. plur. liandyus, 
vrithus, etc. The neuter exhibits no trace of a peculiar sign t 
or m, e. g., faihu. The masculines and neuters preserve the u 
in the singular in the Old High German, but lose the nomina- 
tive s, e. g., sunu, vihu, etc. In the plural the i^-stems pass into 
the Osteins ; and in the Aliddle High German they altogether 
disappear, the masculines and feminines becoming confounded 
with the z-stems, and the neuters with the a-stems. 

I have already mentioned that the primitive distinction between 
the a-, i-, and i^-stems was very much obscured in the case of ad- 
jectives; and that, with the exception of traces, the ^-sterns had 
wholly died out. The z^-forms of the adjective, which were not 
very numerous, took s in the nominative of both the masculines 
and feminines, but the neuters had no sign of gender, e. g.: nom. 
masc. and fern, hardus; neut. hardu. The w-forms died out in 
the Old High German, leaving for all adjectives only a-stems. 

§ 6. Of Consonantal Stems. 

Pure Stems, s-stems. — greek and latin. — 6 juvg, (inus.) gen. 
fivog, which stands for f±v<j-og,=muris for mus-is. In the forms 
like ovg, gen. &)T-6g — (pug, gen. (faor-og, etc., either the r has be- 
come g, or the nominative s has inorganically affixed itself, in 
which case the t dropped out. In either case these forms belong 
primitively to dental tenuis-stems, and not to the s-stems. Mus, 
Jlos, mas, without the nominative sign. Except in vds, vdsis, s be- 
comes r in Latin in the oblique cases, as it stands between vowels. 
It sometimes appears duplicated, as in os, ossis, but here it stands for 
st (compare oariov). 

28 Introduction. 


— greek and latin: semi-vowels — nix for nigv-s, bos for bov-s, 
etc. ; l-stems — aX-e, sal; n-stems — pig for piv-g , the liquid having 
dropped out; (pp{)v, without the nominative sign; Pan, without 
the nominative suffix; r-stems — x £ 'P' ^ ? /p> etc -' ft ir > without the 
nominative signs. 


<p\£\p for (p\if3-g, the -g being the nominative suffix ; urb-s, scobs; 
d-stems — irovg for irod-g ; pes for ped-s; vas for vads, the dental 
having dropped out ; g-stems — (p\6E, for (p\6y-g ; lex for leg-s, rex 
for reg-s. 


yv\p for yvir-g ; ops, etc. ; t-stems — <£we for (j>wr-g ; dens for dent-s, 
pons for ponts, etc. ; h-stems — Xu-yJ for Xv^k-c, <7<£)'j? for o-0/jk-c ; 
^>a# for pac-s. 


rpix-Qi fo'lK for fiiix-Q- 

gothic. — It has been shown m a previous section, that pure 
consonantal stems, properly so called, do not exist in the Gothic, 12 
and that the forms which at first sight might come in here, belong 
rather to the vocalic middle forms, under which they have accord- 
ingly been treated. I shall merely give here a few examples of 
forms which might otherwise have come under the respective 
categories above given for the Greek and Latin: saivs, fraiv; 
bagms, liilm; stols,mel; stiur,figgvs; stabs, lamb; sands, land; 
hugs, gagg; hup-s, skip; shufts, beist; striks, leik; munths, etc. 

Consonantal Middle Forms. — The nominative of some of 
the forms which come under this head exhibit the complete stem, 
which in the oblique case may be unrecognizable, owing to let- 
ter-changes or the dropping of letters. In most cases, however, 
the stem can be better determined from the oblique cases, in 
consequence of the nominative s, or the change of the vowel of 
the affixed syllable so altering the appearance of the stem in the 
nominative as to render it unrecognizable. The form of the stem 
to which the case-endings in the oblique cases are affixed is 
usually called the Thema, to distinguish it from the true stem- 
form, with which it sometimes coincides, but generally not. 
The neuter form of adjectives is best adapted for determining 
their stems. 

s-stems. — In studying the stems of this class, we should be 
careful to distinguish the s-stems proper from words with the 
auslaut s, in some of which the s is secondary, being formed by 

12 Perhaps baurgs (f), a castle, town, gen. sing. nom. plur. baurgs; Gatk (n. 
m.) gen. Guths, and some besides, are exceptions. 

On Roots, Stems, and Derivatives. 29 

the softening of a t, etc., and in others it is the nominative 5, be- 
fore which the liquid n and the mutes d and t have dropped out. 

greek. — Neuters of the third declension in oc ( = Sanskrit as) 
which show the pure stem in the nominative ; in the oblique 
cases the o becomes e, and the g drops out, e. g. — yiv-og, gen. 
ytv-s-og for yiv-za-og, and contracted to yiv-ovg. Adjectival 
substantives in rig, sog = ovg, e.g., r) rpnjorjc ; — forms of this kind 
may be considered as true derivatives. Adjectives in r\g, Eg, 
e.g.: <ra(j>Y}g, vcKpig, gen. aa^-i-og for vaty-ia-og, and contracted to 

latin. — To this category belong certain isolated masculine 
and feminine substantives in Os, such as, honos, arbos, the s of 
which was afterwards softened to r. The adjective vetus comes 
under this head also. The substantives in is and us — pulv-is, 
cin-is, Ven-iis, tell-us, are most probably r-stems, in which the r 
has dropped out before the nominative s. Neuters of the third 
declension in us (= Greek og), the affixed syllable us being 
weakened before the oblique case-endings to or or er, e.g. : 
corp-us, gen. corp-us-is, weakened to corp-or-is, genus, gen. 
gen-us-is, weakened to gen-er-is. 

stems with sonant auslauts. — The steins which come under 
this category are : in the Greek those in v, p ; in the Latin and the 
Gothic I, n, r. M does not occur as the auslaut of a stem in either 
the Greek or the Latin. The pure stem is preserved in the nomi- 
native in the neuter, — the vowel being always short in the Greek. 
The other genders are distinguished in the Greek either by the 
nominative s, before which the liquid drops out, or especially in 
the feminines, by lengthening the vowel of the formational or 
affixed syllable. No such distinction of gender occurs in the 
Latin, the nominative s having given way to the liquid in almost 
every case, except in a very few instances, e.g., sanguis for san- 

L-stems: stem-forming syllable it — masc. Latin pugil, mugil. 

N-stems: stem-forming syllable an — Greek neuter adjective 
juiXav ; an — masc. iraiav, gen. iraiavog ; en — \ijur)v, gen. Xijuiv- 
og, en — -"EXXrjv, gen. (l EX\r}vog ; in softened to en in the nomi- 
native in peeten, and in the derivational suffix of verbal nouns, 
-men, gen. -minis, e.g. — lumen, flumen, etc.; In — clktlv for clkt'iv- 
g; on — Greek adjective iriirov, masc. substantives Sai/uwv, gen. 
Sai/uiovog ', on — \eijuu)v, gen. Xsiiuiovog. To the preceding may 
be added the nouns with vocalic auslaut, which are considered 
to have thrown off the n, and which I have already discussed, 
as, homo, Macedo, carbo, etc. 

R-stems : stem-forming syllable ar — vitcrap, Latin Caesar, 
Gothic Kaisar, fadar; ar — calcar, gen. calcaris; er — 6 ario g en - 

30 Introduction. 

aipog, Latin anser; er—tcpanip — in this and similar words the 
stem-forming syllable may be considered to be Tt)p, and to be a 
derivational one for verbal nouns ; or — pijrojp, gen. prjTopog — 
here the stem-forming suffix is rop, which may be compared 
with the Latin ones in tor and sor, e.g., lector, cursor; — marmor 
is produced, however, by duplication and not by suffix; ur — 
masc. augur, gen. auguris, in which the u remains unchanged in 
the genitive case ; turtur is a stem also formed by duplication ; 
neuters which retain the u in the oblique cases — sulfur and the 
duplicated stem, murmur; neuters which soften the u to o — 
femur, gen. femoris, etc. 

stems with medial AUSLAUTS. — Stem-form i rig syllables : ib — 
adjective caelebs, gen. caelibis; ilb — 6 \aXv^, for yaXvfig, gen. 
XaXvfiog; del — Xajunrag, gen. XajuiraSog, lampas, gen. lampddis; 
ed — merces, gen. mercedis; id — iX-irlg, gen. kX-niSog, cuspis, gen. 
cuspidis, praeses, gen. praesidis ; id — Kpijirig, gen. KpnTrlcog; od — - 
custos, gen. custodis; ud — palus, gen. paludis. Ag is not found 
either in the Greek or Latin; eg — lelex, gen. lelegis; ig — remex, 
gen. remigis; ug — —ripv^, for irripvy-g (in the Greek the nomina- 
tive s fuses with the labial mute b and in the Greek and Latin 
with the palatals), gen. irripvyog. 

stems with tenuis auslauts. — Stem-forming syllables: dp 
— r) XcuXaip for \aT\cnr-g, gen. XalXairog ; ip — adeps, gen. adipis. 
Princeps and similar words do not come here, as they are tine 
compound words in which one of the constituent steins is the 
pure stem ceps. Op and dp occur only in stems forming consti- 
tuents of compound words, e. g., kvkXu)^/, gen. KvuXioirog, etc. 
At — a great number of the Greek forms in at throw off the t in 
the nominative, and are, therefore, somewhat analogous to the 
Latin w-stems homo, ordo, etc., which throw off the n, e.g., aCjfia, 
oro/uLa, Spajua, irpayiia, etc., which form their genitive in roc- 
Sometimes r is replaced in the nominative by p or g, e.g., rjirap, 
gen. r\7rarog ; Kpsag, gen. Kpiarog. To the same category belong 
such forms in it, as jmiXi, gen. juzXirog. The Latin forms which 
may be referred to stems in at, at, et, and et, drop the t in the 
nominative, but retain the s, e.g., anas, libertas, teges (the e be- 
comes long after a vowel, as in abies), quies. So likewise the 
Greek forms in et and it, such as : lading, gen. eaOiJTog ; x^9 L ^-> 
etc. The Latin forms in it have the i softened to e, e.g., miles, 
gen. militis. The following forms also occur: it — Samnis, plur. 
Samnltes; ot — 'ipwg, gen. zptorog; nepos, gen. nepotis; ut — salus, 
gen. salutis. 

To this category belong also the Greek forms in k and the 
Latin in c, of which it will only be necessary to mention a very 
few. Stem-forming syllables : qk, ac — iriva% for rrlvaK-g (we may 

On Roots, Stems, and Derivatives. 31 

also add here the forms in -a/cr, as ava*, gen. avaaroQ) ; ah, dc — 
OwpaZ; fornax, and the adjectives having the derivational suffix 
ac, such as audax, capax, which inorganically retain the nomi- 
native s in the neuter; ek, ec — aXwirr)^, gen., a\u)7T£Kog, the 
neuter lialec, or, fused with the nominative s, masc, halex; ik, w 
and Ic — (potviZ,, salix, gen., salTcis, radix, gen., radicis; 6c — Cap- 
padox; oc,ferox; vk, uc — Kfjpus, gen., KripvKog, Pollux. 

There are also in the Greek stems in vr, v6 but not in v$ ; in 
the Gothic there are also stems in n, (t), and ?id, but as 
my object is rather to show what stems are, than to give a de- 
tailed account of all their forms, I will not dwell further upon 
this part of the subject. 

§.7. Of Derivation. 

Having so often spoken of derivation as distinguished from 
middle forms, and ya-stems, I think it will not be out of place 
if I say a few additional words upon the subject here. Deriva- 
tives are words formed by the addition of affixes to verbal, 
nominal, and other stems. The affixes employed for this pur- 
pose are of two kinds : 1 . Affixes consisting of single letters or 
syllables, which in their present state are not only not inde- 
pendent words, but cannot even be traced up with certainty to 
independent words, though having a definite symbolical significa- 
tion which modifies the meaning of the stem. 2. Syllabic af- 
fixes which afford evidence of their having been once indepen- 
dent words, but which in process of time have been modified 
and have lost that character. 

I have already remarked that Stem-formation cannot always 
be absolutely distinguished from Derivation; this is especially 
true in the case of the stems called middle forms, and derivatives 
formed by the derivational affixes of the first kind, which often 
consist of only a single letter. In discussing the different kinds 
of stems, I have pointed out some examples of this difficulty in 
the case of the Greek nouns in rrjc, the verbal nouns in rop, rr\p, 
tor, sor, and men, for which I proposed the term Derivative Stems, 
that is pseudo-stems formed upon already-existing stems, and not 
starting from roots, as all true stems do. The derivatives formed 
by the second kind of affixes are much less liable to be confounded 
with true stems ; they often have, indeed, almost the character of 
compound words, that is, of words formed by the union of two 
or more stems. The proper distinction between Stem-forma- 
tion and Derivation will, however, be best understood from a 
few examples of the different kinds of words which are formed 
by the latter process. From one kind of verbal form we may 
derive several others, thus, by the addition of the suffixes (Gr.) 

32 Introduction. 

cnc, (Lat.) sc, we get inchoative verbs, as, j3ock:cl> from /3ow, cresco 
from, creo; by trie suffixes (Lat.) it, etc. (N.H.G.) er, etc., we get 
frequentatives, as, cogito from cogo, Mapper n from Happen; by trie 
suffixes (Lat.) #£, ul, etc. (N.H.G.) ^ 5 we obtain diminutives, as, 
scribillo from scribo, ustulo from im), nstum, sduseln from sansen; 
by tlie (Lat.) suffix ess we get intensives, as, capessere from 
capere; by tlie (Lat) suffix wW, we get desideratives, as, esurio 
from edo, esum. Or we may derive verbs from nouns by tlie ad- 
dition of such suffixes as (Gr.) a, &v, aiv, etc. (Lat.) are, ere, ire, 
etc., e.g., Anraw from Xiirag, icoXaictvii) from icoXat,, Xevkciivw from 
Xzvicog, nominare from nomen, lucere from Zw«#, finire from finis. 
We may in turn derive nouns from verbs, thus by the addition 
of the suffixes (Gr.) evg, ™?C? rwp, fxog, etc. (Lat.) tor, tio or ti-on, 
etc. (N.H.G.) el, ung, ing, t, d, etc., we get substantives such as 
ypa(f)£vg from ypafyu), ttolyittiq from 7roi£(x>, 'Vijrwp from piw, 
hvfjfiog from Svld ; victor from vinco, actio from a^o; Hebel from 
lieben, Reibung from reiben, Findling from finclen, Macht from 
7no gen, Jagd from jag en; and by the addition of the suffixes 
(Lat.) ac, 5z7fs, fe, etc., we get adjectives, as loquax (for loquac-s) 
from loquor, placabUis from placo , facilis from facio. So in like 
manner we may get different kinds of substantives from one kind, 
such as diminutives, feminines, etc. ; adjectives from substantives, 
and the converse; adverbs from adjectives, etc., of which, how- 
ever, we need not give examples. 

The greater number of the affixes mentioned in the preceding 
examples belong to the first kind. Those of the second class, 
being, on the other hand, of greater phonetic dimensions, have been 
less intimately fused with the stem, and consequently their histo- 
rical development out of independent words can be more clearly 
traced. This kind of derivation was originally without doubt 
simple composition of the same kind as that by which compound 
words are still formed in living languages. It is the first stage 
of amalgamation from the mere agglutination which takes place 
in the formation of such words, as, penknife, moonshine, etc. 
Its transitional character is made still more evident by the cir- 
cumstance that the affixes of this class are prefixes as well as 
suffixes, and that the former differs from particle composition in 
this only, that in the latter, two independent words still existing in 
the language, combine together, while in the former, an indepen- 
dent stem combines with a letter or stem not now independent. 

In the Greek and Latin the derivatives of the second class are 
neither so well marked nor so numerous as in the Germanic lan- 
guages. The suffixes -aSrjc, -cpopog, -fex, -dicus, etc., are really 
stems, and consequently we may consider w^ords ending in them 
to be compound words, rather than derivatives, e. g., OsotiSfjg, 

On Roots, Stems, and Derivatives. 33 

KavntyopoQ, artifex, mendicus, etc- In the English we have a 
number of well marked derivational suffixes of this class; e.g., 
-7i00^ = N.H.G. -heit, Goth, haidus, way, condition, as for instance, 
girlhood; -sAzp=N.H.G. -schaft, O.H.G. scaf, shape, property, 
etc., as partnership ; -c?om=N.H.G. -thum, Goth, dom, primitively, 
judgment, tribunal, dignity or condition of a person in general, 
as, for instance, dukedom; -some, a stem which signifies similarity, 
and, hence, Goth, sama, Eng. same, e.g., handsome; -fo/ = N.H.G. 
-lich, Goth, leiks, O.H.G. licit, Eng. like, similar, equal. Compare 
in the Romance languages the Italian suffix -mente, Fr. -ment 
(e.g., sainement, purement), from the Lat. mens. 

§ 8. Of Composition. 
Composition is the union of two or more stems, or even words 
with grammatical endings, so as to form one word, and may be 
looked upon as the highest stage of word-formation. Some lan- 
guages possess the power of forming compound words with great 
facility, especially the Greek and Sanskrit. Among modern 
languages, German possesses it to some extent. Two kinds of 
Composition may be distinguished, the Synthetical and Parathe- 
tical. The first kind is where the first word loses its inflection, 
that is, occurs as a stem, and the last alone is inflected; the 
second kind consists of mere juxtaposition, each element of the 
compound retaining its inflexion. The parathetical may be con- 
sidered to be the first stage of composition. Particle composi- 
tion, such as that by which componnd verbs are formed by pre- 
fixing prepositions, comes under the category of parathetical 
composition In the older language-periods a copulative vowel 
was frequently introduced between the constituent words — a 
phenomenon which offers a remarkable analogy to the stem 
copulative vowel. In the Greek, this vowel was generally o, 
seldomer i, or e; in the Latin i, and exceptionally o, or u; in the 
Old High German it was generally a, afterwards e; and in the 
Modern German, as in the English, it has dropped out, 13 or an s, 
and in the former language an en, which are flexional endings, 
have taken its place, e. g., ^/^(o^po^oc, carn(i)fex, nacht(i)gall, 
Hulf(s)buch, Tasch(en)buch, doom(s)day. It is worthy of remark 
that the English word night(iii)gale presents a kind of transition 
between the simple copulative i and the more usual Modern Ger- 
man en. The copulative vowel belonged, in the older languages, 
only to noun forms, and not to those obtained by the union of 
verbs and particles. Combination is sometimes accompanied by 
phonetic changes in one or both of the constituents ; such, for ex- 
ample, as that which takes place in the stem-vowel in the Latin 

13 It is, however, sometimes retained in N. H.G., as in Tage-buch. 

34 Introduction. 

verbs, legere, colligere, and which has been already noticed when 

discussing the subject of progressive assimilation, etc. 

One of the constituents of a compound word represents the 
fundamental idea or basis of the conception; the other, the 
secondary idea by which the former is determined, modified, or 
limited. The former may be compared to the root of a word. 
and the latter to the grammatical affixes : with this difference, 
however, that the latter are chiefly suffixes, while in compound 
words the fundamental word is usually the last member ; the qua- 
lifying word is consequently prefixed, eg., bride-grc: i. al.iss-wbi- 
dow, and window-glass. In some Greek verbal nouns the reverse 
position of the constituent members is apparent, e.g., <hi\6\oM:,c, 
etc. It was probably the oldest form of composition, but has 
almost wholly disappeared from written language, even from the 
Sanskrit. Curiously enough, it exists both in the spoken English, 
French, and German, e.g., breakfast, tire-botte, taugenichts. This 
circumstance is interesting as to the question of the origin oi 



§. 1. TJie Accusative Singular, 

As the classification of stems discussed in the foregoing chapter 
is based upon the manner in which they become nouns by 
afrixing the nominative sign, I was obliged so far to anticipate 
the subject of flexional endings, as to describe in section 4 of 
the preceding chapter the character of the nominative ending. 
I need not, thereibre, say anything further on that point here, 
and will accordingly pass on to the oblique cases, and first to the 
Accusative Sincmlar. 

The sign of the Accusative in Sanskrit, Zend, and Latin, is m; 
in Greek v, Lithuanian and Old Prussian n. It is probable that 
in the primitive Indo-European language it was likewise m. 

£AHK The m was affixed: 1. directlv to vocalic stems of 
the masculine and feminine forms of substantives and adjectives — 
via-m, jide-m, cive-m, manu^m; 2. with an intercalated copulative 
to all consonantal sterns — reg-e-m, arbor-e-m. 

The consonantal stems which have passed over into apparent 
vocalic stems, alluded to at p. 24, follow the rule of consonantal 
stems in the oblique cases, that is, require a copulative : ration- 
e-m, carbon-e-m. According to some philologists, the i- (e-) 
stems also take the copulative vowel like consonantal sterns, the 

Case- Endings of Xuuns. 35 

declension vowel, or stem vowel, giving way before the flexional 
copulative. According to this view, civem would be civ-e-m, not 
cive-m with the i of the stem changed to e as was assumed above. 
The first view is the simpler and more rational. 

The Sanskrit m is usually transformed by anusvdra li into the 
nasal n. The Lithuanian n is also similarly weakened. In the 
Latin the in was generally disregarded in prosody, and suffered 
elision before vowel anlauts. It was dropped altogether in the 
most ancient Roman inscriptions, as, for example, in the epitaph 
of L. Cornelius Scipio, who was consul a.u.c. 494: Hec cepit 
Corsica Aleriaque urbe x% for Corsicam Aleriarnque urbern. It 
is curious that in the modern romance languages the nominative 
singular has been frequently formed from such mutilated accusa- 
tive forms: Italian — buono, imperative, leone, — bonum, impera- 
tcrem, leonem. The Portuguese on the other hand retains in 
many instances the m — liomem, virgem, som,z=hominem virgi- 
tiem, sonum. The Italian forms its nominative plural from the 
corresponding Latin case — parte, servi, -=.porto3, servi; the Span- 
ish, on the contrary, forms it from the accusative plural — ricos 
hombres, los servos, los caballeros. The Oscan has preserved the 
accusative m in all declensions. 

greek. The Latin declensions are richer and more varied 
than those of the Greek. In the former there are five, which, 
however, may be reduced to three ; the fourth may be included 
under the third, and the fifth under the first, by which we can 
assimilate them to the Greek. 16 The fuller endings of the Latin, 
as, for example, the plural ones (-rum, -bus, etc.), may perhaps 
be attributed to the absence of the article, which gives such 
lucidity to the Greek declension, while it helps to weaken it, by 
rendering the endings less indispensible, and perhaps also to the 
frequent use of prepositions in the place of a greater number of 
cases. The v may be found directly affixed to the vocalic stems 
as in the Latin: ypav-v, irnyy-v- The stems in w and ev are, 
however, an exception, as they do not form their accusative in 
v : 77^w, r)\6-a ; fiauiXsog, fia<n\l-a. Attention has been already 
drawn to the anomalous character of the stems in w (p. 24 j, 
which Curtius believed to be relics of n- stems. According to 
the hypothesis of Heyse regarding civem, mentioned above, these 
stems would be considered to take a copulative, before which 
the declension vowel o> gave way, and that afterwards the v was 

1 * Anu-svara,° or "after sound", is the term used by Sanskrit grammarians 
for the marks . (n) that is, a weakening of a nasal auslaut. 

15 Bunsen — Beschreibung der Stadt Bom. III., 6L6, sqq. 

16 A system which has been very successfully followed by Dr. Donaldson in 
his Latin Grammar. 


36 Introduction. 

dropped; so that the primitive form of ri\o-a would have been 
i)y6-av. Many other explanations may also be given: it does 
not, however, come within the objects of this introduction to 
discuss them. 

gothic. The accusative sign has been wholly lost in the 
Gothic, except in the masculines of the adjectives, so that the 
accusative form of substantives presents us with the naked stem. 
The n in the accusative forms belonging to the' weak declension, 
such as hanan, tuggon, etc., belongs to the stem, but was dropped 
in the nominative, by which a class of apparent vocalic stems was 
produced, to which allusion was made at p. 24. In the mascu- 
lines of adjectives, we find the accusative sign preserved in the 
form na, the a being merely an inorganic addition, which was 
dropped in O. H. German, while the n has been preserved in 
N. H. German : Goth, blinda-na, N. H. G. blinde-n. 

§.2. The Genitive Singular. 

Sanskrit and zend. The genitive singular endings in the 
Sanskrit are: masc. and neut., sya; masc. and fern, s; masc., 
fern., and neut., as and fern. as. In masc. and fern, the endings 
s or as may be considered to be practically the same, the former 
being affixed to vocalic stems, and the latter to consonantal; 
especially as the stem vowels in the i- and w-stems are always 
gunated in the genitive: e.g., kavi-s, sunu-s, gen. Jcave-s=:kavai-s, 
sunos=sun-au-s. With these endings the feminine ending as- of 
vocalic stems naturally connects itself, because if the stem vowel 
be short, the genitive may be formed by s alone with a gunation 
of the stem vowel, as well as with the ending as: e.g., prit-es = 
prit-ai-s, or prit-y-ds In the latter the stem vowel has been 
changed into y ; when the stem vowel is long, the % u are inva- 
riably changed to y, v, and after a-stems a y is added, so that 
the endings are in reality -yds, vds. The genitive singular end- 
ings in Zend are: hS (also hyd) = Skr. sya; ao = Skr. as; s = Skr. 
s; and o = Skr. as. 

latin. The whole of the a-stems, that is those declined 
according to the first, second, and fifth declensions, no longer 
form their genitives singular in s. The word paterfamilias = 
paterfamilias has, however, preserved the true ancient form of the 
a-stems of the first declension, which corresponded with those of 
the same declension in the Greek. And, again, on old monuments 
we still find suaes provinciaes=zsuo2 provincial. The genitive 
ending of the first declension has thus become m by the loss of 
the s after the diphthongation of the stem vowel, In the second 
and fifth, declensions the genitive ending has been replaced by 
an affixed i, which had probably originally a locative significa- 

Case-Endings of Nouns. 37 

tion ; in the second declension the flexional i absorbs the stem 
vowel — scamn-i; in the fifth declension the stem vowel is not 
absorbed, and except that after a consonant it is shortened, it is 
not further affected — die-i, fide-i. 

Stems with consonantal auslaut and pure vocalic stems, that is, 
all nouns of the third declension, with the exception of the 
middle forms in i (e), affix s with a copulative i, corresponding 
to Skr. a, Gr. o (ogrrLat. is) : gru-is, urb-is. The observation 
made respecting the ^-sterns, when discussing the accusative 
ending, explains the reason why the i-stems are excepted; 
some philologists believing that they take a copulative in the 
genitive also. The zj-stems of the fourth declension belong like- 
wise to this category ; we have the old forms fructu-is, senatu-is, 
afterwards the s dropped off' and the ui contracted to u or i, as in 
the dative : senatu. According to the oldest inscriptions, as for 
example the Senatus Consultum de Bacchanalibus, it would ap- 
pear that the copulative of the genitive was not i, but o or u, as 
in the words nomin-us, senatu-os, domu-os, and later domu-us. 

Bopp traces the genitive ending ins of some pronouns and ad- 
jectives to the Sanskrit genitive ending sya. He supposes jus to 
be obtained by displacement from sya or sja: hu-jus, cu-jus, illi- 
us for illi-jus, etc. Donaldson, on the other hand, looks upon 
the Latin jus as a weakened form of the ending yds. May not 
this latter form represent in fact the first modification, which, 
according to Bopp's view, sya must have undergone ? In con- 
nection with the latter view it may be mentioned that Steinthal 
has made the ingenious suggestion that the primitive genitive 
suffix was sya, which he considers to be made up of the nomina- 
tive s and the relative pronoun stem ya (fern, yd), so that we 
might have two forms, a masc. sya and a fern, syd; the latter of 
which would give exactly the fern, suffix yds, while the Latin 
jus might have come from the masc. sya. 

In the Oscan the genitive singular ending was as, for the first 
declension, and eis for the second and third: Djuv-eis = Lat. 
Jov-is. Here the Oscan forms are fuller and richer than those 
of the Latin, for besides preserving the s in all cases, we have 
traces both of the stem and the copulative vowels in the second and 
third declensions, while the former has been absorbed in the Latin 
second declension. In the Umbrian the genitive ended in s. or r. u 

greek. The genitive singular is formed in the Greek by : 

(«) Affixing g to the feminines of the first declension in a, 17, 
the inorganically shortened a of the nominative becoming a 
or 7), corresponding to the Sanskrit feminine vocalic stems which 

17 See the paradigms of the Umbrian declension quoted from Aufrecht u. 
Kirchoff' s, Sprachdenkmaler, p. 115 sqq. in Donaldson's Varroniamis. 


38 Introduction. 

take the ending as — Movva, nupa, gen. Movarrj-g, ireipa-g. The 
Attic ending wg of the z-stems is considered by some as the 
complete representative of this Sanskrit as, by which ir6Xewg = 
iroXyog is compared with the Skr. prityds. But, as Ebel points 
out (p. 83), the Homeric TroXrjog leads rather to iroXzyog. 
The ending tog is not confined to the feminines, for we have the 
masculine fiacriXi-wg. 

(h) By affixing g with a copulative o (Skr. a) to stems with 
consonantal auslaut, pure vocalic stems and vocalic middle forms 
in i, v, to), ev : \up-6g, crw/zar-oe, Kt-6g, aXriOi-og, l)(9v-og r)\6-og. 

(c) Many nouns do not form their genitive in g ,as for exam- 
ple the a stems of the first declension in rig, ag, and those of the 
second declension in which the primitive a has passed into o ; in 
the Attic these nouns have ov in the genitive. The Attic ov of 
the first declension was obtained like the Ionic ew and the Doric 
a, from the Homeric ao (fiop£ao,Alveiao) which was obtained from 
a-io, and this from a-aio — Skr. sya, by dropping a. Bopp like- 
wise explains the ov of the second declension from sya; in the 
stems in a, and in the pronouns of the third person, d-sya be- 
comes o-(tlo, the a then dropped out by which the Epic o-io 
was formed, and then oo contracted to ov ; thus Xoyoio, Xvkoio, 
and the Old Epic toio must have been obtained from the older 
forms: Xoyo-aio; Xvko-<jio = Skr. vrXka-sya; to-gio = Skr. ta-sya. 

goteic. The Gothic z'-stems which exist in the syncopated 
form in the nominative, affix the genitive s to the full form, the 
stem vowel i of the feminines being gunated : masc. nom. synco- 
pated from gast-s (full organic form gasti-s), gen. gasti-s; fern, 
nom syncopated form ded-s (for dedi-s), gen. dedai-s, nom. syn- 
copated form anst-s (for ansti-s), gen. anstai-s. A similar guna- 
tion occurs in both masc. and fern, of w-stems, which likewise 
directly affix s to the gunated stem : mas. sunu-s, gen. sunau-s; 
fern, handu-s, gen. handau-s. The mas. and neut. of the a- and 
?/a-stems affix the gen. s to the stem by a copulative, i, which 
replaces the declension vowel a, or, in other words, they have 
passed over to the i-declension : jisk-s, dag-s, vaurd, gen. jishi-s, 
dagi-s, vaurdi-s; nom. and gen. haryis, hairdeis. The masculines 
of ya-stems which decline according to the strong declension, 
are therefore the same in the genitive as in the nominative 
The feminine a-stems, on the other hand, have preserved the 
declension a in the oblique cases — in the genitive as o, but 
strengthened however before the genitive s: giba, gibo-s, a form 
which may be compared with the Skr. genitives in as. 

The masculine and neuter substantives and adjectives of the 
weak declension affix the genitive s directly to the n, which is 
the universal ending of the bases belonging to the weak declen- 

Case-Endings of Nouns. 39 

sion: masc. hana, hanin-s ; neut. hairto, hairtin-s; fern, tuggo, 
tuggon-s; fern, managei, managein-s. 

In the Germanic languages the genitive s has been preserved 
in all the strong masc. and neut. ; but the fern, already lost it in 
the O. H. G. ; and in the N.H.G. they have lost all the flexional 
endings in the singular. The copulative vowel i of the masc. 
and neut. a and z-stems, becomes throughout e in the O.H.G: 
Goth., masc fiskis, O.H.G. , visces; Goth., neut. vaurdis, O.H.G. 
ivortes; Goth., masc. gastis, O.H.G. gastes. The fern, of the 
a-stems lose the s in the O.H.G., but retain the long vowel, 
which, however, becomes short in the M. H. G. : Goth., giba, 
gibos ; O.H.G, kepa, kepa (0); M.H.G., nom. and gen. gebe. 
The fern, of the z-stems likewise lose the s in the O.H.G. : 
ansts, anstais; O.H.G., anst, ensti. The declension vowel of 
the masc. w-stems likewise becomes e — Goth., sunus, sunaus; 
O.H.G., sunn, sunes ; but the feminines appear to pass into the 
^•-declension, with the loss of the genitive s — Goth., handus, 
handaus; O.H.G., hant, henti In the weak declension, the 
genitive s is lost in the O.H.G.: Goth., masc. liana, hatiins; 
O.H.G., liano, lianin ; M.H.G., liase, Jiasen ; neuter Goth., 
hairto, hairtins ; O.H.G., herza, herzin; M.H.G., herze, herzen. 

Lithuanian and Slavonian. In Lithuanian the genitive sin- 
gular ending is s. The masc. a-stems have lost the s and end 
in 6; according to Bopp tins is merely the lengthened stem- 
vowel which replaces the suppressed case-ending. Schleicher on 
the other hand explains this o as a contraction from aya which 
arose from a-sya. The Lettish has also lost the ending in the 
corresponding stems, while the Old Prussian has preserved it: 
Skr. deva-sy a = Lith. diwo, Lett, deewa, O.Pr. deiwa-s. In z- 
and zi-stems the stem vowel is gunated as in Gothic, and the 
Lithuanian has preserved the guna in the masc. as well as in the 
fern, z-stems; as in Sanskrit the ai is, however, contracted to e: 
Lith. aiu^-s = Skr. ave-s. The genitive s has been lost in Old 
Slavonic ; consonantal stems end in e, o-stems have the primitive 
a of the stem, w-stems u, and z-stems the naked thema. The 
Sanskrit ending sya is, however, represented by the pronominal 
ending go: SI. to-go = Skr. ta-sya. 

§ 3. The Dative, Locative, and Instrumental, Singular. 

Sanskrit and zend. The singular dative endings in the San- 
skrit are: mas., fern., and neut., S; fern., ai; masc. and neut. 
a-stems, ay a. In Zend the endings are also e, and ai. In the 
Greek and the Latin it was perhaps i in all declensions. The 
dative ending in the Gothic was perhaps a (e). The singular 
locative endings in the Sanskrit are: masc, fern., and neut., i; 

40 Introduction. 

fern. am. The masc. u- and i- stems, and sometimes the fern. 
also, have a peculiar locative in au, before which the stem-vowel 
is dropped, or becomes y. Bopp supposes that it Was obtained 
from as, and that it is, therefore, a genitive form used in a loca- 
tive sense. The singular instrumental endings in the Sanskrit 
are: masc, fern., and neut., a, (yd); masc. and neut. a-stems, a, 
with an intercalated n: e. g., givS-n-a. The Sanskrit locative 
endings i and dm are represented in Zend by -i, and -a? and the 
instrumental by a. 

latin. — In the a-stems of the first declension, the datrve i, 
instead of producing ai, fuses with the stem vowel into ce. The 
o of the second declension, like the corresponding Greek w, has 
arisen from oi, as is proved by the old datives, popoloi Romanoi. 
It is worthy of remark, that while the locative i suppresses the 
thema vowel in the genitive, the latter, in most cases, absorbs 
the former in the dative. In the third, fourth, and fifth declen- 
sions, the % is affixed as an independent sound, and often even 
inorganically lengthened : su-i, urb-i, fructu-i. In some forms of 
the fourth and fifth declension, the i is suppressed by the thema 
vowel, u, e : tactu (Plautus), usu (Lucretius), facie (Lucilius). 
The genitive ending having been replaced by a locative i in the 
fifth declension, the genitive and dative coincide in that declension. 

In the Oscan, the dative of the first declension was formed in 
ai, like the locative, which had also 03; in the second declension 
the dative ending was ui, the locative being in ei, and in the 
third declension the dative ending was ei, corresponding to the 
Umbrian in e for all declensions, which, unlike the ablative, was 
probably (at least originally), long, although the morte in the epi- 
taph of Plautus — Postquam morte datu st Plautus Comwdia luget, 
if it be an Umbrian dative := morti, is short. The Oscan and Um- 
brian dative endings ai, ei, and e, obviously lead to the charac- 
teristic dative endings ai, 3, in the Sanskrit. This would seem to 
show that the Latin dative i may not have had originally a loca- 
tive signification, but is a true descendant of a primitive dative. 

greek. — The Greek dative i fuses with the thema vowels, a, 
e, 0, into a false diphthong in the a-stems, that is in the first and 
second declensions: v)fi£pq, vlicy, o'/ko), for which we have also 
the form o'Ikoi. With the vowels £, o, a true diphthong is pro- 
duced : noXei, ri^ol. The stems declining according to the third 
declension affix the dative i directly to the stem, without modi- 
fication. The Epic form, <j)i, of the dative ending, will be no- 
ticed in the section on the dative plural. The dative suffix in 
the Greek, and perhaps also in the Latin, appears to have had 
originally a locative signification, and which several words still 
show : & g , 2aXa/xTvt, MapaOuvi. 

Case-Endings of Nouns. 41 

gothic. The mas. and neut. substantives of the a-, ya-, and 
z-stems belonging to the strong declension, form their datives in 
a; in the a- and ya-stems the ending coincides with the stem 
vowel, and in the i-stems replaces it: fiska, liarya, hairdya, 
gasta. In the O.H.G. the dative of these forms is also «, 
which in M. and N.H.G. becomes e. The feminine a- and i- 
stems form their dative in i, which is however gunated as in the 
genitive : giba, gibai; deds, dedai. We might explain gibai with 
Bopp to be for gibai-a, with diphthongation of the stem vowel, 
the dative sign having fallen off. In the same way the O.H.G. 
dative of geba, gebo (u) would likewise be an extension of the 
stem vowel without a dative sign. In the case of the fern, i- 
stems, we must suppose upon Bopp's view that the stem vowel 
was gunated : dedai for dedai-a. The dative of the z/-stems may 
be explained in the same way ; there is no proper dative ending, 
but instead of it au, produced by a gunation of the sterna, as in 
the genitive: siui-au, hand-cm, for sun-av-a, hand-av-a. In the 
O.H.G. the gunating a is replaced by an i. The consonantal 
middle forms have lost their dative sign : fiyand for Jiyanda; 
brothr for brothra. The nouns of the weak declension have no 
dative suffix ; in the masc. and neut. they all end in in; and in 
the fern, in on and ein; these endings become masc. in, fern, un 
and in in the O.H.G., and en in all genders in the N.H.G. 

According to Bopp the dative sign a was originally the suffix 
of the instrumental = Skr. a. The masc. and neut. a- and z-stems 
of the O.H.G. substantives, and adjectives, belonging to the 
strong declension show an instrumental in u: tagu, gastu, wortu, 
which Bopp believes to have arisen from a. The original instru- 
mental has thus assumed a dative meaning, while a phonetically 
different form has been developed out of it to express the instru- 
mental. In the Gothic the adjectives have in their strong declen- 
sion a special dative ending for the masc. and neut. ; the feminines 
on the other hand correspond with substantives ; in the O. and 
M.H.G., however, the feminines have likewise a peculiar ending. 
These endings are pronoun endings which have passed over to 
the adjectives. The following paradigm will render this passage 
obvious : 

Masc. and Neut. Fern. 

r - 








Goth. . . 

. blind-amma . . . 

. th-amma 



O.H.G. , 

, . plint-emu (emo) . 

. d-emu 



M.H.G. . 

. d-em 



Lithuanian and Slavonian. The dative ending in Lithuanian 
is i (in fern, z-stems ei). In Old Slavonian consonantal and 
w-stems end in i. This t, according to Bopp, corresponds to the 

42 Introduction. 

Sanskrit dative ending e — ai. masc. and neut. o-stems end in u; 
fern, a-stems in e; and Masc. and fern, ^-sterns in i. In Lithu- 
anian the locative ends in e and je. Although this e is short, 
Bopp thinks it has arisen from ai produced by the stem-vowel 
and the locative i. In O. Slavonian the locative ending is i in 
consonantal and w-stems, and is therefore apparently identical 
with the dative ; in masc. and neut. osteins, and fern, a-stems it 
is L The locative i has been lost in Lettish, the stem-vowel is 
however lengthened in a-stems. The instrumental ending in 
Lithuanian is mi, which is evidently connected with the plural 
instrumental ending mis = Skr. bins, Zd. his. Masc. and fern, a- 
stems do not, however, take the ending, the former end in u f 
and the latter coincide with the nominative. 

§. 4. Tfie Ablative Singular. 

The Sanskrit, Zend, and Latin, have an ablative singular, but 
in the dual and plural they express the ablative signification by 
the dative, as other languages do by the genitive In the 
Sanskrit the ablative endings are: masc. and neut. a-stems t; 
masc, fern., and neut. of the other declensions as, which 
resembles the genitive. But, as Bopp has concluded from the 
analogy of the first and second personal pronouns, mat, tvat, and 
from the Zend ablatives, the primitive ablative form was t. This 
is further confirmed by the ablative suffix in the oldest Latin, 
and in the Oscan, being d, and therefore quite distinct from the 
dative. Thus we have on the Columna Rostrata: presented 
sumod Dictator ed olorom in altod marid pucgnad vicet. ls The d 
was however frequently apokoped : e. g. mari for marid, senatu 
for senatud, etc. In the Umbrian the ablative ends in a vowel 
which is sometimes a, i, and u, as well as e, and therefore does 
not always correspond to the dative. In the Sanskrit the abla- 
tive has the signification of whence in the sense, of space; in the 
Latin it has, however, a wider application, because, in addition 
to the proper ablative meaning, it often combines in its applica- 
tion a locative and instrumental signification. 

The ablative sign may also be recognized in adverbs, as in 
bened, facillumed, which are evidently the old ablatives, bonod, 
or bonud, facillumod or facillumud; and in prepositions suprad, 
entrdd. According to Bopp, the ablative sign is also found in the 
enclitic pronoun met (z=Skr. mat from me), which occurs in the 
compounds egomet, memet, and in the conjunction sed, anciently 
written set. The suffix tus ( = Skr. tas) in ccelitus, and the de in 

inde, unde, are perhaps likewise related to it. 


18 Donaldson's Varronianus, 2nd Ed., p. 229. The Oscan form praesentid 
occurs on the Bantine Table, 1. 21. 

Case-Endings of Nouns. 43 

§.5. Tlie Dual. 

Peculiar dual forms of substantives are only to be found in the 
Sanskrit, Zend, Greek, Slavonian, and Lithuanian. 19 In the San- 
skrit the dual endings are: nom., ace., and voc. niasc. and fern. 
au (in the Veda dialect a, and in the Zend a) ; neut. % (which 
fuses to e with the stem a); dative, instrumental, and ablative 
bhydin in all genders ; genitive 6s, yds in all genders ; and loca- 
tive 6s, which coincides, therefore, with the genitive. The Greek 
has only two dual forms: 1. that of the nominative, accusative, 
and vocative; and, 2. that of the genitive and dative. The nom. 
ace. and voc. dual sign, in all three genders of the third declen- 
sion in the Greek, is £ ; in the first and second declensions the 
ending is suppressed and the stem vowels a, o, lengthened to a, 
u) ; the same thing takes place in the masc. and feminine of the 
i- and w-stems in the Sanskrit. The dative (also genitive) dual 
suffix in the Greek is iv, which in the first and second declension 
is affixed after the stem vowel, which fuses with the i into a 
diphthong; in the third declension the affix is oiv instead of lv. 

The Old Slavonian has preserved the dual ending more com- 
pletely than the Lithuanian. Masc. a-stems end in the nom. 
ace. and voc. in a; fem. and neut. a-stems in e (ye) ; masc. and 
fem. z-stems in -i; and z«-stems in ii (generally = Skr. u), or they 
pass into the a declension ; consonantal stems in i (sometimes in 
e). The genitive and locative end in Old Slavonian in all gen- 
ders in u, and the dative and instrumental in ma. In Lithuanian 
masc. a-stems end in the nom. ace. and voc. in u, which Bopp 
explains to have arisen from the Veda ending a. In the i- and 
w-stems the ending is suppressed, as in the Sanskrit, but the stem 
vowel is not, however, lengthened, as in that language. The 
Lithuanian genitive dual ending ft, is borrowed from the plural ; 
Bopp now however thinks that this u is a true dual ending, and, 
like the corresponding O. Slavonian u, connected with the 
Sanskrit genitive dual ending os: Lith. divej-u = Sk.T. dvay-ds 
duorum, duarum (see Dr. Ebel's opposite opinion, p. 84). The 
locative is lost in Lithuanian. The dative and instrumental 
dual in Lithuanian is m. 

If we compare the dative dual suffix in the Sanskrit bhydm, 
which is also that of the instrumental and ablative, with the 
plural suffix for the dative and ablative bhyas = Lat. bus, the y 
being ejected, we see that bhya is common to both, and may 
therefore be considered as the proper dative suffix, while the 
proper dual sign may be assumed to be m, and that of the 

19 See "On the Celtic Dual", §. 10. p. 85, for Ebel's observations on the 
relics of dual forms in Irish. 

44 Introduction. 

plural s. The dual sign m has been lost in the Zend, and the 
dative accordingly ends in bya. Respecting the Greek dative 
dual suffix there are two hypotheses. Bopp believes tv to be 
nothing more than a crippled form of bhyam. Diintzer, on the 
other hand, believes that the dual sign v=fi is simply affixed to 
the singular dative form in i. 

The only traces of special dual forms in the Latin are duo 
and ambo; and in the Gothic and Old High German they are 
only to be met with in the personal pronouns. 

§.6. TJie Nominative and Vocative Plural. 

The nominative and vocative are alike in all Indo-European 
languages. In the Sanskrit the masc. and fern, ended in as: the 
neuters ended in i, which was affixed to the stem with an eupho- 
nical n between the stem vowel, which was lengthened, and the i. 
In the Zend the masc. and fern, ended in 6, which represents 
the Sanskrit as, and the nom. and ace. plur. of neuter nouns in 
a, which was also the ending in the majority of the old languages 
of the family. Bopp considers the Sanskrit i as merely a 
weakened form of such an a. In the Oscan the first declension 
ended in as, and the second in us; and in the Urnbrian, besides 
as and us, the endings ar, or, also occur, the r being obviously 
formed from s. In a fragment of Pomponius, which is found 
in Nonius Marcellus, we meet with the nom. plur. laetitias 
insperatas. These forms perfectly represent the Sanskrit as. 
The Greek eg and the Latin es of the third declension likewise 
represent the Sanskrit ending. In the fourth declension in the 
Latin the ending is us, the u being formed by the fusion and 
contraction of the stem vowel with the e (a) of the ending. In 
the fifth Latin declension the ending es results from a similar 
contraction of e-es into es. We may explain, in the same 
manner, the long e of the ending of the i-stems of the third 
declension, in which it is organic, as a contraction of i-es. In 
the other forms of the third declension the e is inorganically long. 
The duplicated form ds-as, which occurs in the Vedas, and which 
appears to have been intended to mark in a very material manner 
the plural number, has been suggested as an explanation of this 
inorganic long e in the Latin ; but the simplest explanation is to 
suppose an invasion of the form of the z-stems. This tendency in 
the Latin to give z-fonns to nouns of the third declension, which 
had not them originally, is illustrated by such words as navis from 
vavg, civis from the Oscan cevs, etc. 

All the masc. and fern, a-stems of the first and second declen- 
sions in the Greek and Latin end in i. In the Greek the i com- 
bines with the stem vowels a, o into ai, 01 ; in the Latin a- forms 

Case-Endings of Nouns. 45 

of the first declension, the stem vowel and ending combine to ai, 
which, as in the genitive, passes into oe. We have evidence 
of this passage in the Senatus Consultum de Bacchanalibus, 
where we find tabelai datai for tabelloe datoe. The i of the case- 
ending has absorbed the u of the a-stems of the second Latin 
declension in us — popidi, domini; bnt in Old Latin it was poplce, 
from poplo-i, etc. On old inscriptions we find, instead of i, the 
anomalous nom. plnr. ending eis = is: hisce magistreis. 

Neuter nouns form their nom. and ace. plural in a, which in 
the a- stems of the second declension in the Greek and Latin is 
affixed in place of the stem vowel o, it, which is dropped — Swp-a, 
dona; in the third declension the a is affixed without dropping 
the declension vowel — 'iSpi-a, mari-a; this is also the case in the 
fourth Latin declension — cornu-a. 

According to Bopp the plural ending as is merely " an 
extension of the singular nominative sign s, so that there lies in 
the extension of the case suffix a symbolical indication of plu- 
rality". This seems to imply that the a of the ending is in reality 
the plural sign. This affords a simple explanation of the circum- 
stance that, the nom. ace. and voc. plur. of neuter nouns are 
formed by dropping the nominative s, which has a certain 
positive gender character. Grimm and other philologists believe 
the true sign of the plural to be s. Many forcible reasons may 
be given in support of this view, which, however, cannot find a 
place here. The plural ending i may be looked upon as the 
pronominal ending, Skr. £ = a primitive ai, Lith. and Goth, ai, 
Gr. 01, Slav, i, which invaded the substantives. 

gothic. All masculines and feminines both of the weak and 
strong declensions end in s in the Gothic. The masc. and fern, 
a-stems of the strong declension end in 6s, which represents the 
Sanskrit as, the long vowel of the Gothic being the result of a 
contraction of the stem and case-ending vowels. The i-stems 
end in eis; the zt-stems in yus (for ius). The ending of the masc. 
and fern, of the weak declension is ns, which is directly affixed 
to the stem ; the n is the characteristic sign of the weak declen- 
sion, and, as has been pointed out at p. 25, is added to all the 
cases except the nom. sing, and dat. plur. All the neuters end in 
a; those of the weak declension having the characteristic n before 
the a. The following paradigm will illustrate these rules : — 

Strong declension. 


Sing. Plur. Sing. Plur. Sing. Plur. 

Masc. fisks . . . fiskos balgs . . . balgeis sunus . . . sunyus 

Fern, giba . . . gibos ansts . . . ansteis handus . . handyus 

Neut. vaurd . . . vaurda • faihu . . . 

46 Introduction. 

Weak declension. 

Masc. hana . . hanans 

Fern, tuggo . . tuggons managei . manageins 

Neut. hairto . . hairtona 

In the O.H.G. the s dropped off in every case ; in the strong 
declension the long vowel remains : vise, vised; pelk, pelhi; sunu, 
sunt. In the weak declension the stem-forming a, 6 of the Gothic 
has been obscured to u, and the ending is accordingly un : zunga, 
zungiin. All the neuters drop the a, so that those belonging to 
the strong declension have no ending, while those belonging to 
the weak end, like the masc. and fern., in un: daz wort, diu wort; 
herza, herzun. 

In the M. and N.H.G. all the different vowels of the ending 
become e in the masculines and feminines, so that the strong 
nominative plurals all end in e, and the weak in en. This e has 
likewise invaded the neuters in the N.H.G., which as a rule take 
e : Worte, Sometimes instead of e, they take up er. This suffix 
is not a flexional ending, and does not exist in the Gothic; it 
first made its appearance, according to J. Grimm, in the O.H.G. 
as ir attached to neuters. In the N.H.G. it has however invaded 
the masculines also, in which, as well as in the neuters, the root 
vowel is frequently diphthongated : neut. worter; masc. manner, 
geister. The neuters of the weak declension end in M.H.G. like 
the weak masculines in en : herze, Jierzen. 

The plural forms of adjectives declining according to the weak 
declension in the Gothic, Old and Middle High German, are like 
those of the substantives. In the strong declension, on the other 
hand, they have, with the exception of the Gothic neuters, forms, 
which like those of the dative singular, appear to have passed 
over from the pronouns. The following paradigm will illustrate 
this invasion of the pronominal endings : 




Goth, . . 
O.H.G. . 
M.H.G. . 

r ■ i 
Adject. Pron. 

. . plint-e. . . di-e 
. . blind-e . . di-e 

r " *> 
Adject. Pron. 

blind-6s . . th-6s 

plint 6 . . . di-6 

blind-e . . . di-e 

Adject. Pron, 
blind-a . . th-6 
plint-u . . di-u 
blind-iu. . d-iu 

It is worthy of remark that the modern languages, — Spanish, 
Portuguese, French, and, with few exceptions, English, — form 
the plural of all nouns in s. 

Lithuanian and Slavonian. The s of the Sanskrit ending as 
has been preserved in the Lithuanian ; the masc a-stems have, 
however, taken the pronominal ending, which in substantives is 
the diphthong ai, and in adjectives i. The s of the ending as 
has been lost in O. Slavonian, but the vowel has been pre- 
served as e. The crippling of the diphthong ai to i, which 

Case-Endings of Nouns. 47 

occurs in Lithuanian adjectives, extends to siibstantives and pro- 
nouns in O. Slavonian: vliiki lupi, for vluJcoi, ti=h\, om = illi. 
The Lithuanian, on the other hand, contracts ai to e in the pro- 
nominal declension: Lith. <e = Lat. hi, Skr. U, Goth, thai, Gr. rot'. 
In Old Prussian, substantives, pronouns, and even adjectives of 
masc. a-stems have ai, or occasionally for it ei and oi. In Lithua- 
nian the stem-vowel is lengthened in i- and zj-stems ; in Sanskrit 
the stem-vowel is gunated in the corresponding stems : Lith. dwys- 
z=Skr. avay-as; Lith. svnu-s=: Skr. sunav-as. In Gothic evidence 
of a similar gunation is found in the endings of the i- and u -stems 
(p. 45); the gunating vowel has been preserved as i, which in 
^-sterns fuses with the stem- vowel to i (ei), and in i^-stems becomes 
y before u: gastei-s, sunyu-s for suniu-s. 

§. 7. The Accusative Plural. 

The accusative plural endings in the Sanskrit are — masc. and 
fern, s, as; masc. n; neut. i. In the Zend these endings are: 
6 ( = Skr. as) for masc. and fern, consonantal, %-, and ij-stems 
which is affixed with or without guna; o (Skr. s) fern, a-stems; 
s ( = Skr. s) fern, i-, and t«-stems; and the peculiar ending eus of 
masc. and fern, nouns in r, which Bopp explains from ahs, the 
n becoming vocalized, and the a changed to e. In the Oscan, 
the first declension had ass, and the second ziss. The masc. and 
fern, a-stems of the first declension in the Greek and Latin 
ended in as — Movtrag, mensds; those of the second declension 
in ovg in the Greek, and os in the Latin. The i-stems, and the 
stems with consonantal auslaut of the third declension in the 
Greek, have the ending ag, which, in the case of the latter, is 
affixed to the pure stem or thema : iroSa, TroSag. The w-stems, 
which retain the u in their thema, end in Greek in vg: nom. 
plur. l\0ve g, ace. l\9vg. The a-stems of the fifth declension, the 
^-sterns, and those with consonantal auslaut of the third, and the 
w-stems of the fourth declension in the Latin, coincide with the 
nominative plural, as do the accusative plural of all neuter nouns 
in the Greek, Latin, etc. In the older Latin, however, the 
accusative plural of z-stems and also of consonantal stems, ended 
in eis or is. In the Umbrian the accusative plural ended iny in 
all declensions. 

The accusative plural ending of all masculine nouns, and of 
the feminine forms of the u- and z-stems in the Gothic, is ns, 
which is directly affixed to the full stem form : fisha-ns, oalgi-ns, 
sunu-ns. The accusative plural of the feminine a-stems has not 
n, and therefore coincides with the nom. plur. : gibos, bidos, etc. 

In the Old High German, the accusative plural coincides 
throughout with the nominative. 

48 Introduction. 

The universality of the ns in the Gothic accusative plural, and 
the circumstance that m (n) is the sign of the accusative, suggests 
a very simple explanation of the plural suffix. Grimm, in tact, 
regards it as the accusative singular -\- the plurals: the primitive 
form would accordingly be -ms (-ns). The sign of the accusative 
has therefore been wholly lost in the plural in the Greek and 
Latin, and, except in the masc. a-stems, in the Sanskrit also; the 
Greek and Latin have preserved the plural s, while the Sanskrit 
forms which have preserved the ?i have lost the plural s. The 
Greek and Latin accusative plural endings must, therefore, have 
dropped an n, so that ag and as stand for arc and ams; ovg for 
ore, and as for urns, oms. 

Some examples of this complete accusative ending have been 
preserved in the Greek dialects, e.g. Tov-g = Tovg. It has like- 
wise been preserved in Old Prussian in the same form as in the 
Gothic, both the masc. and fern, having the masc ending ns; the 
Lithuanian, on the other hand, has only preserved the s? Skr. 
decd-n, O.Pr. Mtva-ns deos, Lith. dekuu-s. In Lithuanian the 
stem-vowel of fern, a-stems, masc. and fern, /-stems, and masc. 
a-stems is short, while in the nom. it is long". The a of masc. 
a-stems has been weakened to u. In Old Slavonian the accusa- 
tive ending has been lost in all masc. and fern, stems ; steins in n 
or r, however, add an i, which must probably be explained by 
a transition into /-stems. 

§. 8. TJie Genitive Plural. 

sanskbit axd zend. The usual ending o£ the genitive plural 
of substantives and adjectives in the Sanskrit is dm, which is 
affixed directly to consonantal steins, and to vocalic stems by 
means of an euphonic n between the stem vowel and that of the 
ending: pad-am; civd-n-dni, priti-n-dm. This dm was probably 
primitive sdm, a form which in fact we find in the pronouns 
which preserve the primitive forms longer and completer than 
the nouns, e.g. in the demonstrative te-shdm, (horum), tdsdm, 
(harum). The s is the sign of the genitive singular, so that am is, 
properly speaking, the genitive plural sign The genitive plural 
in Zend is anm} in the a- and a-stems this ending takes a 
euphonic n, as in the Sanskrit. 

latin. The a-stems of the first, second, and fifth declensions, 
form their genitive plural in rum. This rum represents the 
Sanskrit sdm, and must have been anciently sum, which in 
turn leads to an earlier sam=Skr. sdm. The Oscan genitive 
plural suffix zum appears to confirm this view. The /-stems, and 
the consonantal stems of the third declension, and the w-stems, 
form their gemtive plural usually in wm=Skr. dm: mari-um, 

Case- Endings of Nouns. 49 

lapid-um, fructu-um. The i-um of the i-stems has penetrated into 
many forms among consonantal stems, such as terbium, serpen- 
tiurn, etc. On the other hand, many ?'-stems drop the stem- 
forming -i in the genitive plural: can-urn, vat-um. In some 
antique forms belonging to consonantal stems of the third declen- 
sion preserved in Varro and Charisius, the full form rum is 
affixed to the stem by means of a copulative e: lapid-e-rum 
instead of lapid-um, bov-e-riun instead of bo-um. We may also, 
however, consider them as formed from the genitive singular by 
the addition of urn: bover-wmzzbovis-um ; lapidfr-umzzlapidis- 
urn. While, on the one hand, the full form rumzzsdrn was 
sometimes found in nouns of the third declension, many a-stems 
of the first and second formed their genitive plural in urn: 
agricol-urn, vir-um. 

greek. The genitive plural ends in the Greek in wv = Skr. 
am. The ending of the first declension has a circumflex, 
which points to an original form a-wv : Mouawv, old form 
Movarait)v. This u-iov probably represents a fuller form awvzz 
Skr. sdm = Osc&n zumzzhat. rum, so that Movaawv would re- 
present a still more complete form Moucra-crwv = Lat. Musarum. 
In the second declension the copulative o dropped out before the 
ending : Xoy-ojv. In the third declension the suffix attaches itself 
directly to the thema in consonantal and vocalic stems : 7roc-wv, 
IxOv-(dv, -rrrixz-ojv, p>am\i-bjv. In the stems formed by the stem- 
forming suffixes og and eg, in which the <x drops out, and the 
thema vowel is z, the latter is however contracted with the end- 
ing: Tpinpuv for rpinpi-wv ; aatyCov for aa(pi-wv. 

gothic. In the masc. and neut. forms belonainrr both to the 
strong and weak declension of the a-stems, the genitive plural 
ends in e; that of the feminine a-stems ends in 6; and of the mas- 
culine and feminine i and ?^-stems likewise in e. In the O.H.G. 
all the a-stems form their genitive plural in 6, and the i and u- 
stems in i, o, the usual n being intercalated in the weak declension 
before the ending. The strong feminine a-stems likewise intro- 
duce an euphonic n betw r een the stem and the ending, as in the 
Sanskrit: Goth. fern, gib-d, O.H.G. kepo-n-6 (cf. Skr. r-wd-n-dm). 

In the Middle and New High German, all the strong forms 
end in e, and the weak forms lose the vowel-ending, so that the 
genitive is always the same as the nominative. 

The e, 6 of the Gothic is derived from a, so that the s and m 
of the primitive ending have been wholly lost, and the vowel only 
preserved. The s has however been preserved as z in the Gothic 
in the adjectives and pronouns declined strongly ; in the adjectives 
the stem vowel is diphthongated. The Gothic z becomes r in 
the M. and N.H.G. 

50 Introduction. 

Lithuanian and Slavonian. The Lithuanian genitive plural 
ends in u. The Old Prussian has lost the vowel, and preserved 
the consonant of the ending am as n. It has also preserved the 
full form = Skr. sdm, in its pronominal genitive plural ending 
son. The Old Slavonian has u; in the pronominal declension 
it has, however, the ending chu, which Bopp explains as = Skr. 
sdm, O. Pr. son. 

§. 9. Tlie Dative, Locative, Instrumental, and Ablative Plural. 

The dative and ablative endings in the Sanskrit for all gen- 
ders is bliyas; for the instrumental the ending for all genders is 
bhis, but the masculine a-stems form an instrumental in is. The 
locative plural for all genders is su (shu). In the Zend the dative 
and ablative end in byo, which fully represents the Sanskrit 
bliyas; the instrumental ending is bis; and the locative hva 
( = Skr. su), sva ( = Skr. shu). 

There are two forms of the dative plural ending in the Greek 
and Latin, one of which is considered to be more ancient than 
the other. The older form in the Greek is ai, aiv, and in the 
Latin bus; the newer form, which is alike in both, is is. The 
latter occurs in the first and second declension in the Greek and 
Latin; in the former language the stem vowel combines to a 
diphthong with that of the ending ; in the latter the stem vowel 
fuses and contracts to is — mensis for mensa-is. The Oscan dative 
and ablative plural endings a-is (first declension), u-is, o-is 
(second deck), present us with similar uncontracted forms. The 
suffix is may be looked upon with considerable probability as 
the locative singular joined with the plural sign s. 

The old Greek form <h, oiv, which is found in the third declen- 
sion, originally occurred likewise in the first and second declen- 
sions, as is proved by the old datives fiovaaicri, \6yoiai, a cir- 
cumstance which shows that the ending is is of later origin. In 
the Old Greek we find the fuller form <x<n — TravT-z-acn, Kvv-e-aai ; 
this form occurs not only in Homer, but also in the Eolic, and to 
some extent in the Doric dialects. Aufrecht, Benfey, and others 
consider this suffix a&t to bave arisen from gfl = Zd. sva. The 
a and o of the stems in the first and second declension must 
have changed into at and 01 under the influence of the final i ; 
this lengthening of the vowel in its turn reacted upon the 
ending, and one a dropped out. 

The Latin dative and ablative plural ending bus, which cor- 
responds to the Sanskrit bhyas,m&yhe explained from the dative 
ending bi in tibi (=ztu-bi) sibi, and mihi, the b being softened to 
h in the latter. This bi or blii corresponds to the Sanskrit bhy- 
am, hyam: tu-bhyam=tibi; mahyam mutilated from mabhyam-=. 

Case- Endings of Nouns. 51 

mihi. It had evidently a primitive locative signification. Bopp 
compares it with the Sanskrit preposition abhi, with which 
the German bei (English by in the locative sense) is to be 
connected. We may also connect ibi, which is the locative of 
the pronominal stem i-s, and the analogous form ubi. Aufrecht 
has shown that the basis of the endings bi in ibi and_i«fo\ and im 
in illim, istim is a fim, which may be recognized in the Umbrian 
locative plural suffix fern, which drops the m, and sometimes is 
weakened to /. 20 And further that the Old Epic (pi, (piv is the same 
case suffix. From all this it is evident that bhyas = bus is sim- 
ply a singular locative dative bhi, combined with the plural 
suffix as. Again, the plural ending of the Sanskrit instrumental 
is bhis = Zd. bis =:h&t, suffix in nobis, vobis. Here too we have 
evidently a compound suffix composed of a singular 6/ri=:Lat. 
bi, and a plural s. This Latin suffix fully represents the 
Greek (piv (in $aKpvo-(piv, 6pe<j-(ptv, etc.), which must have been 
originally (pig = bhis, for (pig bears the same relation to (piv, that 
the first person plural suffix fieg among the verbal endings, does 
to the other form fizv. This original (pig was composed of a 
singular (pi and the plural g; this (pi is now used along with the 
plural form (piv without distinction for singular or plural, at one 
time (piv and at another (pi ; the latter was, however, originally 
singular and the former plural. In the Old Latin bus was used 
in the first declension also, and at a later period to distinguish 
the genders — deabus , ftliabus ; and seldomer in the second declen- 
sion — jilibus, amicibus. Bopp thinks that the newer is of these 
declensions has come from abus, obus, which, in the first place, 
became ibis and then is, by dropping the b. Aufrecht, on the 
other hand, believes the is to have arisen from ins. 

Gothic. The sign of the dative plural in the Gothic and 
O. H. G. was m for all nouns. In the M. H. G. the m is replaced 
by n. The ending m w ? as attached directly to the thema in the 
strong declension. In the Goth, and M.H.G. the m of the case 
ending took the place of the characteristic n in the weak declen- 
sion. In the strong adjectives the thema vowel a becomes ai 
in the Goth., and e in the O. H. G. ; in the masc. nouns it is ob- 
scured, and in the fern, it becomes 6. The Gothic dative plural 
m is connected with the Sanskrit and Latin endings, bhyas, bhis, 
and bus, bis, by the corresponding Lithuanian case suffix mus in 
mumuszz nobis, yumus=:vobis, which appears in all other words in 
the syncopated form ms. The Gothic has accordingly softened 
b to m, and dropped the plural s : fiskam for jishams zzpiscibus. 

20 The locative in the Umbrian appears to be formed by the addition of em to 
the accusative singular and plural, thus: ace. sing, tutam loc. sing, tutam-em; 
ace. plur. tutaf. loc. plur. tutaj-em. 


52 Introduction. 

A trace of v the original ms remains in the Old Norse form.3 
tveimr, thrimr — Jjat. duobus, tribus. 

Lithuanian and Slavonian. In Lithuanian the dative plural is 
formed by the addition of the pronominal ending mus or ms 
above-mentioned; masc. a-stems end, however, in is. In Old 
Slavonian the ending is mil, which is evidently a weakened and 
crippled form of mus. The Lithuanian instrumental plural 
ending is mis, which is apparently the singular ending, to which 
the plural s is added. In the Old Slavonian we have also this 
ending in the crippled form mi — the final consonant being 
generally lost in that language. Stems in o form their instru- 
mental in %, in which Bopp recognizes the Sanskrit d-is, Zd. d-is 
(in which the b has been lost), Lith. a-is; according to this, the 
s was lost and the i produced umlaut of the stem-vowel. Masc. 
and neut. yo- (ya-) stems form their instrumental in i. The 
locative plural is formed in Lithuanian by the endings sa, su, or 
se, or, as in the Lettish, more frequently by s only. In Old 
Slavonic the locative plural is formed by the pronominal genitive 
ending chu. 

The annexed paradigm, which contains all the case-endings 
mentioned in the preceding pages, may be found useful in com- 
paring the relative state of preservation of each case-ending in 
the different languages of the family. 


Nom. and Voc. -s, m, -t . . 

-m . . . . 






-sya, -s, -as, -a 

-e, -ai, -aya . 

-i, -dm, e, au 
-a (-ya), n-a 
-t, -as . . . 

Zend. Latin. 

(-s), -m, t . -s, -m, -d. 
-m, -hm, -Sin -m . . . . 

-he, -In/a, do -is, -us, -i, -<B 
(-s),-6. . 

-d, -ad 

(•i (fused to a) 

-s. (-m), -d 

(for oi), -u, -e, e-i 

a-i, a-e, e-i 



-s, (-m). 

-(/(when lost the same a-d, u-d, i-d, -a, -e, -i, -u . 
as the dat., except that 
a is represented by &.) 



-£, -v. . . 

-s, -ta . . . . 

-V, -a . . . 

Lost except in masc. 
adj., where it is -na 

-£, -og, -ov . 

- s 

i (sometimes 
fused into £, 
y, V ) . .' . 

-a, -ai, -au 

O.B.German. Lithuanian. Old Slavonian. Old Prussian. 

Lost (-r, -z) . . . -s . . 

Lost except in masc. -h, -n 
adj., where it is -71 

s, when lost -a, -i, -n -s, -o . 

-a, -i, -iu, -in, -in, -un -i, -ei 

-c, je 

(-'<•), lost . 

Nom. Ace. and) -au (in the Vedas -a) 
Voc. J -t 

Genitive and 



-os, yos . 


-uo, -a, -i 



-f, -T, v 

-a, e, t, u 




-6, -a 

(-as), -re, - 
cs, -a, us 

J -as, -lis . 

its, -d (and other -ass. -uss? 
forms like thi] nom.) -j'ss 

■ rum, -ttm 

-zum, -urn 
-ium, -im 1 


' -os. -eis, -yus 
~) -ns, -na 

-ar, -ag, -ovg -ns, -os 

ls < ■"' \ -d,-i,-un 

-6, -o n-o, n-o 


■ai (ei, oi) 


-bhyas . 


-bhyas . 


-bhis, -is 


-su (shu) 

-by 6 . . 

-bis, -is 
-hva, -sva 

{-bus, -is . . . -ais, -uis, -ois (- 

\ iss 1 . •< 


-}MS, -is . . . 
-sa, -se, -su, -s 






§. 1 . Bopps view of the aspirations and eclipses in Modern Irish, 
and the modifications which it undergoes through the Old Irish 

BOPPS sagacity has never been, perhaps, so brilliantly proved, 
as in the discovery that the whole of the aspirations and eclip- 
ses, by which the Modern Irish declension is apparently disfigured, 
are nothing else than the relics and results of the after-action 
of the old case-endings. 21 Zeuss 1 determination of the old forms 
of the article has confirmed this supposition in the most complete 
manner, as regards the n and the consonant aspirations; the t 
and h before vowels are, however, to be somewhat differently ex- 
plained. After what Zeuss remarks (pp. 59 and 63), 22 we cannot 
help regarding the h as, in the beginning, a useless and arbitrary 

21 Die Celtischen Sprachen, etc. S. '22, et seq. 

22 ( a ) [The passages in Zeuss are as follows : — 

P. 59 : "H is not found as a radical in the Irish ; and if in ancient MSS. S 
besides the combinations ph, th, ch, the h is also seen alone, which only happens 
at the commencement of words, it is nothing more than a breathing prefixed 
to the initial vowel, as in the ancient Gaulish names : Hercynia, Helvii. This 
h, neither a radical nor a necessary letter, occurs, without any fixed rule, in 
one place, and is not found in another ; as : uile, huile (all), Wb. fq. ; €ula 
(wise), Wb., heulas (wisdom), Sg. 209 a ; aui, hdui (descendants), Sg. 28 a 30 b ; 
and so on. The ancient language knows nothing of that regular usage accord- 
ing to which the modern dialects, Irish and Gaelic, prefix the h in a hiatus to the 
initial vowel of a substantive following the forms of the article na (gen. sing, 
fern., and nom. and dat. plur.) or preposition ending in a vowel. We find, 
indeed, for example, inna kirise (of the faith), Sg. 209 b , but also inna idbairte 
(of the offering), inna indocbale (of the glory), inna anme (of the soul), na 
cecilse (of the Church), Wb. 22° 22 b 25 c 27 a ; na accobra (the desires), Wb. 20 c ; 
la Atacu (with the men of Attica), Sg. 147 a ; a oentu (from unity;, Wb. 26 b 
a albain (from Scotland), Marian. Scot. ap. Pertz. 7, 481". 

P. 63, " The s drops out by ' infection' in the ancient language.* The more 
recent language, indeed, which expresses the aspirate in its primary state as a 
strong s, almost as ss, pronounces the same letter when mortified or ' infected' as h , 
but I think this h is of still more recent origin than the A in a hiatus between the 
article or a preposition and the initial vowel of a substantive following, of which 
supra. For the ancient Irish MSS. either mark the mortified s, like the^ by a 
dot [the punctum delens, used commonly in mediaeval MSS. to mark a letter 
written by mistake, and to be omitted], or else omit it altogether".] 

*[" Infection", or "mortification", as it is called by some grammarians. Dr. O'Donovan 
calls it "Aspiration"; which he defines thus: "Aspiration, a grammatical accident, the 
general use of which distinguishes the Irish, Gaelic, and other cognate dialects of the Celtic, 
from all other modern languages, may be denned as the changing of the radical sounds of the' 
consonants from being stops of the breath to a sibilance, or, from a stronger to a weaker 
sibilance".— O'D., Gram., p. 39-40. 


56 EbeVs Celtic Studies. 

addition before vowel anlauts, which, at a later period, perma- 
nently fixed itself after vowel auslauts ; for the passage of s into h 
appears to be foreign to the Gaedhelic branch of the Celtic ; in 
the dative plural, where h likewise appears before vowels, it is not 
s, but b, which has dropped off; for from do?iabis' 23 the Modern 
Irish dona has been first developed through the Old Irish donaib 
or donab. On the other hand, we also frequently find the t 
(Zeuss, 55, 231, etc.) 24 after n in Old Irish, where otherwise d 
should stand, before eclipsed s. Hence, we cannot look upon the 
t in the nominative of the article as a substitute for s, but must 
assume that it had been prefixed to the s in the more ancient 
forms of the nominative, and afterwards remained when s dropped 
off. The Old Umbrian appears to afford a parallel to this: 
it never shows an ns, except instead of nns in Palsans, but 
either nz or z (enze=onse) or s (neirkabas). Accordingly, in 
the modern form of the lan'guage, this t is to be found wherever a 
vowel has dropped out from between n and s, equally whether 
the s belongs to the article (as in masc. an tiasg, the fish, instead of 
(an(t)s iasg), or the noun substantive (as in fern, an tslat, the 
rod, 25 instead of an(t) slat). 2& It is absent when s or another con- 

23 [i. Kead *dunnabo (from *du-sannabo). Ebel's hypothetical donabis is due to 
his theory that the 0. Ir. dat. plur. sprang from an instrumental (Sansk. -bhis). 
But this theory is destroyed by the Gaulish inscription of Nismes (Revue Archeo- 
logique, 1858, p. 44), in which Dr. Siegfried has recognized two datives plur. — 
viz., matrebo namausikdbo (matribus nemausicis), which are genuine descendants 
of the Indo-European datives plur. in -bhias, Sanskr. -bhyas, the i (j/) being 
ejected as in Lat. -bus. In donaib the stem-vowel a has been weakened into «&'.] 

24 (P) [The passages in Zeuss are as follows : — 

P. 55. "The form NT, also, occurs in forms of pronouns coalescing with 
the preposition in, but only when the preposition governs the accusative 
case: inte (== in earn, fern.) Sg. b a ; intesi (gl. in ipsam) Sg. 199 a , 209 b ; 
intiu (in eos) Sg. 7 a , Ml. 21 a . 28 a . Thence we might expect for the 
other persons the forms : intium (= in me), intiunn (== in nos), intit (== 
in te), intib (== in vos), intis (== in eum), which I have not met with in 
MSS. The harder form, int, of the article prevails before vowels in the sing, 
nom. of the masc. gender, in which, after the usual form of the article, m, 
the hard form of the consonants is retained. Therefore the harder form nt seems 
to contain in itself the signification of action (motion, in the preposition), of 
hardness of form and of the masculine gender ; the softer nd that of the passive 
(rest, in the preposition) of softening [of the letter] and of the feminine gender. 
It is to be observed in addition, that the form of the article int prevails almost 
always (the form ind is very rarely found) before the softened, or, as it is called, 
the ' mortified' s in all the cases of the three genders in which ind occurs before 
vowels (e.g. in the Article) ; this is, however, to be compared with the fact, 
that even the particle ind in composition (in the ancient Gaulish ande-) be- 
comes int before a softened or mortified s in the following word.] 

P. 231. [Gen. sing, of the article, in]. "In, aspirating, before tenues and 
medials ; Ind before liquids, mortified consonants, and vowels. ***** 
Instead of the regular Ind the form INT also prevails before the mortified s, as 
before at p. 55 [extract, supra, note ( 24 )], and here : intsechtaigtha (gloss : "simu- 
ationis"), Ml. 31 a ; fomam intsommai (under a rich man's yoke), Ml. 27 d .] 

25 The difference between an tiasg and an tslat is only graphic, as it is pro- 

On Declension in Irish. 57 

sonant has dropped off; consequently, in the gen. and nom. plur. 
fern, na slaite, instead of nds slaite, in the nom. sing. masc. an 
sruth, the scholar, instead of an(t)s sruth; in the gen. plur. of both 
genders na sruth, na slat, instead of nan sruth, nan slat; in the 
dat. plur., dona srothabh, dona slataibh. 

A third point in which Bopp's view undergoes a modification 
through the Old Irish forms, is the explanation of the nom. plur. 
masc, which in the Modern Irish is formed as in the fern, in na 
with h before vowels, and without alteration of the following con- 
sonants. Bopp thence concludes that in the Celtic the article, like 
the substantive, in the masc. plur. originally ended in as; conse- 
quently, that na has been deformed out of anas; but the Old Irish 
ind, or in with an aspiration following, together with the fern, masc, 
inna or na, show us that here also the masc originally ended in a 
vowel as in almost all the Indo-European languages ; consequently, 
that the modern na owes its existence to an inorganic extension 
of the accusative form, or fern. plur. form, which we already find 
in the Old Irish neutral plural inna, which leaves the consonants 
following unaffected. 27 

With the exception of these three points, the old forms confirm 
throughout Bopp's discovery, according to which the nomin. 
sing, masc, the gen. sing., and the nom. pi. fern., from their very 
origin ended in s; the gen. plur. in n; the gen. and dat. masc, and 
nom. and dat. fern, sing., in vowels. 

The finding of the neuter, which has disappeared without 
leaving a trace in the New Celtic (an or a in the nominative and 
accusative singular, and the plural like the genitive), and of the 
accusative (replaced in the Modern Irish by the nominative), in 
the Old Irish forms inn (before consonants in) in the masculine and 
feminine singular, inna or na in the plural of all three genders, — 
and in which we can still plainly recognize the original ending 
-n in the singular, and -s in the plural, — is an important enrich- 
ment of Celtic grammar. 

I hazard no supposition as to the relation of the old forms with 
i, followed by nt, nd, nn, to the new with a and simple n; the 
vowels of the endings can only be determined through a compa- 
rison of the substantive-declension, to which we shall now proceed. 
§. 2. Stems which belong to the several orders and series of Zeuss. 

The philologist recognizes at first sight, in the first order 

nounced an tlat, as in accurate writing even in the Old Irish s is provided with a 
dot or left out, not only in this case but also after vowels. — Zeuss, 63. 

26 [ii. The t in the nom. sing. masc. of the Irish article has been since shown 
to be due to the law, pursuant to which, in Old Irish, d becomes t before aspi- 
rated s, an tiasg, in O. Ir. mft'asc=an Old Celtic san(d)as-\-escas, subsequently 

27 [See infra. §. 11. On the Article in Modern Irish, p. 88.] 

6 B 

58 EbeVs Celtic Studies. 

(Ordo Prior) of Zeuss (App. I., p. 169), a vocali c(or a conso- 
nantal changed into a vocalic) declension, in the second order 
(Ordo Posterior), consonantal stems; among the latter, the mas- 
culine and the feminine n-stems and nouns of relationship in 
-thir (=r Sanskrit -tar) being especially evident, as had been 
already recognized and put forward by Pictet and Bopp. On the 
other hand, I cannot, from external and internal grounds, agree 
with both these masters in the distribution of the vocalic-stems. 

Namely, if we compare the first paradigm or table of Zeuss 
with the second, his remark, that the first is external, and the 
second internal inflexion, is at once seen to be incorrect. We 
have only to take, instead of ball, a word with e — as, for example, 
fer, man — in order to at once see that the declension of cele (com- 
panion) does not at all differ in the main from that of fer, except 
that in the former a vowel preceded the dropped off ending, in 
the latter a consonant. The vowel of the original penultimate 
undergoes in both the same changes: nom. and ace. sing, and 
gen. plur. cele, like fer, gen. and voc. sing, and nom. plur. celi, 
as fir, dat. sing, celiu &s fur, ace. plur. celiu &sfiru; it is only in 
the dative plural that a slight difference occurs between celib and 
feraib. In short, I. is only a variety of II., and both are related 
to one another, like the Gothic haryis or hairdeis to fishs. Let 
us, therefore, assume for a moment that I. contains z/a-stems, II., 
a-stems; there remain for III. u and z-stems. But a similar 
relation to that between I. and II. also occurs in the feminine 
between IV. and V., and the differences in the paradigm between 
tuare and rainne in the genitive singular, tuari and ranna in 
the nom. and ace. plur., are compensated by the secondary forms 
of the fifth, which we find under the examples gen. sing, -a 
and -o, nom. and ace. plur. -e and -z\ We could here also assume 
in the fourth z/«-stems, in the fifth a-stems, and have only to 
determine then what has become of the i or z-stems, in order 
to remove the objection which could be raised upon external 
grounds against such a division; for, if feminine zz-stems are 
wanting, there is nothing remarkable in the circumstance. We 
shall again find the feminine z-stems under V. ; the z-stems have, 
however, either become ya or z-stems. We find many stems, 
originally consonantal, changed into III. (exactly as in Latin in 
the z-declension) : e.g., dis, dis (aetas) =. Sanskrit ayus, gen. aisso, 
desa. 28 The feminine nem 29 (caelum) = Sanskrit nabhas, gen. nime, 

28 [iii. It is impossible to equate dis with dyus, final s being never retained in 
Irish, not even in the ns- stems.] 

29 [iv. Nem (also nim) was an i-stem — not an a-stem — as we see from the Old 
Irish gen. plur. nime in Oingus Cde D€: — 

Sen a Christ mo labrad " Bless, Christ, my utterance, 

a choimdiu secht nime. O Lord of seven heavens !"] 

On Declension in Irish. 59 

according to V., reminds us of the Slavic forms mentioned in the 
Zeitschrift fur vergleichende Sprachforschung, iv. 342 : voda == 
Sanskrit udan, gora = opog, tima == Sanskrit tamas. If, accordingly, 
we designate the five series set up by Zeuss as : I. b. masculine and 
neutral ?/a-stems ; I. a. masculine and neutral a-stems ; II. mascu- 
line and neutral i and a-steins ; III. b. feminine yd-stems ; III. a. 
feminine a and i-stems, we shall find that this classification will 
receive an external confirmation by a consideration of the words 
and suffixes which belong to the several classes. 

First, most loan-words harmonize, in the most strikingly accu- 
rate manner, with their types in the declension. Compare, for 
example, the a-stems — felsub zzphilosophus, 6rz=. aurum, 30 angel= 
angelus, apstal = apostolus, ejjscop = e^isco^us,fial=. velum, idol 
m. =idolum, ifurnn=infeTmim, saZm=:psaimus, tempul=tem- 
plum, together with the genitive digaim = digammi, metir-=z 
metri; the a-stems — (aZmswi? 31 = eleemosyna, epistil?=.ej)istol&) 
persan = persona, riagol na$ri«Z= regula, pianzz-pcensi, fedb=: 
vidua (no doubt borrowed ?) liter zzlitera, sillab = syllaba ; those in 
ia and id — the masculine notaire, rectaire, tablaire, the feminine 
y^swfo:=:philosophia; those in i — the masculine faith =z vates ; 
in u — the masculine fers — versus, sens = sensus, spirut (gen. 
spirito, spiruto) = sj)iritus. Proper names follow the same rule, 
such as rdm, fem. = Roma, romdn, mas. =Romanus, titzziitus, 
tiamthe = Timotheus, grec — Graacus, although I. a. has here en- 
croached rather more, as the dat. aeneus, gen. adim, Socrait, 
Aristotil, show. 

The forms of the cognate languages afford a further confirma- 
tion, and so do now and then also traditional Gaulish words. Fer 
(sternum) corresponds to the Sanskrit vira still more accurately 
than the Latin vir and Gothic vair (instead of vir stem vira) ; 32 
the adjective fir to the Latin verus (compare rig™ — Latin rex) ; 
din den to the Latin unus ; marb to the Latin mortuus (b = tv) ; 34 
fescor fescar, masc. to the Lithuanian vakaras, Latin vesper; 
bran, raven, to the Slavonian vranu, Lithuanian varnas (San- 
skrit varna) ; run fem. to the Gothic runa; dia masc, the ano- 
maly of which is only apparent, to the Latin Deus (instead of 
Deus = Sanskrit deva) ; fere fem. probably to the Greek opyri ; 
tuath fem. to the Oscan tovto, Umbrian toto; anirn, fem., from 

30 [v. or is even found with the n of the neuter termination in the nom. sing. 
or hglan (pure gold), where or h is exactly the Greek avpov.~\ 

31 [vi. Recte airman : epistil is right — the i'm the last syllable being due to pro- 
gressive assimilation — a phenomenon which Irish exhibits in common with 
Finnish and Magyar.] 

32 The Lithuanian form wyras, and the rarity of the Latin i, instead of a, before 
r, speaks more in favour of vira than of vara. 

33 [vii. Kecte n=Gaulish rix, a #-stem.] 

34 [viii. The b in marb, now marbh, is a v. marb=*marva, Welsh, marw.'] 

60 EbeVs Celtic Studies. 

which several cases are formed according to III. a. (Zeitschrift f. 
vergl. Sp. vi. 213), and which corresponds in these to the Latin 
anima; Ian to the Latin plenus (see supra) ; column to the Latin 
columbaf 5 ardd, probably, to the Latin arduus ; nu m to the Latin 
novus = Sanskrit nava, on the other hand, nue is related to the 
Gothic niujis = Sanskrit navy a; aile to the Latin alius; conse- 
quently we may refer uile to Gothic alls by assimilation from 
Ij: the neuter cride represents exactly the Sanskrit hrdaya (less 
accurately the Greek Kapdia), trede neut. (the Trinity) the San- 
skrit tritaya; muir shows itself by the Gaulish mori- to be an 
z'-stem, which, notwithstanding small deviations, the Latin mare, 
Slav, more, Gothic marei, confirm; mug (servus) appears to be 
identical with the Gothic magus (puer), and consequently an u- 
stem ; fid neut. (arbor) resembles the Old Saxon widu, Anglo- 
Saxon wudu, Old High German ivitu (Old Norse masc. vtf&r), 
and besides is shown by the Gaulish vidu to be a w-stem, like cath 
(pugna), bith (mimdus) by the Gaulish catu-, bitu-; the fern, set 
(dat. seit, pi. seuit) via = Gothic sin/>s, like det (dot. deit = 
Lat. dens), fluctuates between i- and a consonantal declension ; 
finally the double forms ben and ban (rnulier) may be explained 
either from* gvina 31 ( = Gothic qvino) and *gvano ( = Greek yvvri, 
Boeot. fiava) or from *gvani ( = Sanskrit jani) and *gvana (exactly 
as the Slav, zena can have been formed from zana or zina). 

But even if we considered these agreements as merely acci- 
dental, much more would the identity of the suffixes come out. 
The adjectives come almost without exception under the classes I. 
a. and b. in masc. and neut. III. a. and b. in femin., consequently 
to a- and yd-stems, which in all the Indo-Eruopean languages are 
the most numerous. The superlatives end in -em, of which I have 
found no inflexions in Zeuss, and are probably derived from ima, 
or am, certainly from -ama, which is inflected according to I. a. 
Of the adjectives the fern, abstracts in -e are very generally 
formed according to III. b., which corresponds to the Sanskrit yd, 
Lat. -ia, Greek -ia, Old High Germ, -i, Middle High German -e, 
e.g., amprome (improbitas) from amprom, sulbaire (eloquentia) 
from sulber, doire (miseria) from doir, soire (nobilitas) from 
soir, firinne (justitia) from, fir ian, luinde, bitterness, from lond, 
n6ibe (sanctitas) from noib, etc. Among the masc. in -e (I. 
b.) the words in -ire or -aire, corresponding to the Slav, -art, 
as echire, echaire (mulio), and many loan-words (from the Lat. 

35 [ix. Colum (recte colomb), gen. coluimb, is a masc. a-stem, not fern, like 

36 [x. Recte nua. The nom. plur. of sit (see below) has the masc. article in 
Zeuss, p. 237.] 

37 [All words to which an asterisk is prefixed are hypothetical.] 

On Declension in Irish. 61 

-arius) distinguish themselves ; among the adjectives those in -de 
= Sanskrit -tya, only of larger use, e.g. nemde (coelestis), talmande 
(terrestris), colnide (carnalis), etc. ; the Sanskrit -taya occurs in 
the numeral adjectives dede, trede corresponding also in gender 
to the Sanskrit tritaya, catushtaya. We must, therefore, accord- 
ingly compare the modern fern, in -mhuin, as produced from the 
older -maine, not with the Sanskrit neuter in -man, but with the 
Latin fern, in -monia (seachmuin=sechtmaine, consequently not 
accurately corresponding to the % Lat. septimana), especially as 
even the Old Irish already sometimes exhibits retrenchment, as 
testemin, festimin stands by the side of the Lat. testimonium, the 
neut. aill by that of the mas. aile=. alius. 

The verbal substantives, which take the place of the infini- 
tive, are particularly interesting. Those of them that apparently 
contain the naked root, as cumang (posse, potentia), fulang (to- 
lerare), may be recognized by their declension according to I. 
a., as a-stems, to which the Sanskrit gerund in -am, and the 
locative in -e, by which the Indian grammarians frequently ex- 
plain the roots, are parallel. Pictet (De raffinite des Langues 
Celtiques avec le Sanskrit, p. 161) compares the infinitive in t, 
th, d, dli, with the Sanskrit -turn; Bopp (p. 56) rather with the 
Slavonic -ti, especially because of the form tinn; we find among 
the suffixes in Pictet, the Irish adh compared with the Sanskrit 
-atliu. We shall become acquainted with tinn further on under 
consonantal declension ; about the other forms the Old Irish sup- 
plies us with information. There -ad and -ud follow the second 
mode of inflexion, -t the third; we are consequently the more 
entitled to presuppose in the former two suffixes w-stems (like 
the Lat. -tus, from which the supine, Sanskrit -tu, from which 
the infinitive and gerund -turn, -tvd), as, according to the latest 
statements of Schleicher (Beitrage, I. 27), even the Slavonic 
infinitive in -ti belongs to this formation ; on the other hand, the 
feminine forms in -t (according to III. a.) are not to be sepa- 
rated from the feminine abstracts in Sanskrit -ti, Greek -A (oi), 
Lat. -ti (si), Gothic -ti, \i, di. The feminines in -dl (III. a.) 
remind us of the peculiar Slavonic participles in -lu; but it 
would be difficult to decide whether -a or -i has dropped off 
in them. The feminine in -em are a-stems, which correspond 
to the Greek verbal-nouns in -juij ; the masculine in -am, -om, 
-um, remain obscure to me. Finally, -ent, -end, according to I. 
a., I consider to be borrowed, a supposition to which the forms 
legend, scribend, already point. The masculines in -id, gen. 
-ada, in which Zeuss, p. 766, suspected an original -at, still 
deserve to be mentioned ; the proper stem-ending is -ati, abso- 
lutely like the Sanskrit -ti, Greek n (in /udvTig), only differently 

Q2 EbeVs Celtic Studies. 

employed, as it appears in the Irish, as a taddhita suffix. 38 The 
part. perf. pass, appears to be the only exception to this regular 
correspondence with the cognate languages : they do not end in 
-th or -d, according to I. a., as the analogy with the Sanskrit, 
Greek, Latin, and Gothic would lead to, but in -the, according 
to I. b. ; but the original form still lies before us in the preterite 
passive of the impersonal conjugation (sing, -d, plur. -tha); we 
have consequently to recognize in the ordinary form an addition 
{-ya or -aya) similar to that in the Old Welsh -etic. On the 
other hand, the part. fut. pass, -thi, properly -thi, accurately links 
itself to the Sanskrit -tavya, Greek -riog (Lat. -tivus). 

If, finally, we compare the forms of the article, which, accord- 
ing to Bopp's view, also belong to an a-stem, and exactly agree 
with the stems in I. a in the distinguishing cases, gen. sing, and 
nom. pi. masc, there will be found sufficient external grounds to 
justify our division. We shall now pass to the inner character- 
istics which exist in the Irish phonetic relations, in order to de- 
velope and explain, as far as possible, the individual forms. 

§. 3. Test afforded by Irish Phonology for determining induc- 
tively the Primitive Forms of the Celtic Case-Endings. 

The Irish vocal system exhibits two very close points of con- 
tact with the German, the umlaut or obscuring of an a by i 
and u, and the fracture of an i and u by a. In reference to the 
first, it is particularly remarkable that the three kinds of assimi- 
lation of the a before i and u, which we generally find separated 
in different languages and language-periods (complete assimilation 
as in the Sanskrit giri and guru, diphthongation as in the Zend, 
pairi and pauru, umlaut proper as in the Old Norse hendi and 
hond), appear here side by side ; thus the well-known particle ar- 
is written air-, ir-, er- [and aur-'j ; the accusative plural of ball, at 
one time bullu, at another baullu; rolaumur (audeo) also rolomur. 
As umlauts of a there consequently occur: — 1, ai or i, more 
rarely e; 2, au or u, more rarely o; inversely % changes into e, o 
into u, under the influence of an a following, as in the Old High 
German ; thus, for example, in the gen. feda, moga from fid, 
mug. We may see how far the last law has extended itself, from 
the fact that it has even invaded foreign names, as e£aZ= Italia; 
but when Zeuss ascribes the same influence to a succeeding o 
and u, it should be considered that o and a often interchange, as 
in the gen. etha or etho from ith, where the e owes its origin 
rather to the a than to the o; but, on the other hand, o and e 
arise from simple weakening — namely, before double consonants, 

38 [So the Indian grammarians call the secondary suffixes.] 

On Declension in Irish 63 

so probably also in felsub — philosophus. 39 We can just as little 
recognize an umlaut of the e into i, for where we have reason to 
consider e as primitive, there is produced by a succeeding i or u, 
not i or iu, but ei or eu, for example, in the plural geinti (gentes) 
in the dat. neurt, from nert (virtus, valor). 40 Zeuss has pro- 
ceeded in a one-sided manner, inasmuch as he has everywhere 
taken the vowel which appears in the nominative as the primi- 
tive one ; while, in cases like nime, giun, it is rather the i 
changed into e by a that again appears. On the other hand, it 
must be admitted that umlaut is sometimes produced by an e not 
derived from i, as in gen. rainne from rann (pars). According 
to this, the rule for the Old Irish (we pass over here the vowel 
changes in the Modern Irish, and slight deviations, such as oi for 
ai, ea, eo, for ao) may be expressed somewhat as follows : under the 
influence of a succeeding a, i changes itself into e, u into o ; under 
that of a succeeding i (exceptionally also an e), a into i or ai 
(or e), e into ei, u into ui, o into oi; finally, under that of a suc- 
ceeding u, a into u or au (or o), i into iit, e into eu. It is unne- 
cessary to observe that the factor very often disappears, and the 
fact remains, so that, just as in German, we can determine by 
the vowel-changes in the stem the vowel of the ending, a cir- 
cumstance of so much the more importance, because it will soon 
appear that the Irish, even in its oldest form, is much more 
weakened in the auslauts than, for instance, the Gothic. 

If we apply the rule just given to determine the vowels of 
these endings, we obtain, in the first instance, for the masculine 
and neuter, according to I., the following endings : — 

Sing. Nom. -(a)s, -(a)n . . Plur. -i, -d 

Ace. -(a)n . . -u, -d 

Gen. -i . -an 

Dat. -u -abis 41 

Examples: ball (membrum) ball, baill, baull or bull, baill 
ballaib or ballib; fer (vir), fer, fir, fiur, fir, firu, fer, feraib; 

39 [xi. Here (at least as to the breaking of i into e by o) Zeuss seems right 
and Ebel wrong. Thus : 

Bretan=Brito (Book of Armagh) ; lenomnaib (lituris), Zeuss, 739, compare 
Lat. lino ; lebor from liber (Zeuss, 744) ; senod (Cormac), from si/nodus (y=i) 
cenel=ceneth(o")l=01d Welsh cem'tol.] 

40 [xii. E seems changed into i by a succeeding i in the following instances : — 
Aristotj'l (gen. sing.), Zeuss, 887, magisttr, nom. pi. of magister, Zeuss, 1057, 
heriti'c (=haereto'ci) Zeuss, 1055.] 

41 [xiii. Regarding the remarks in notes 23 and 39, the hypothetical endings 
for the masc. and neut. may be set down as follows: 

Sing. nom. . . . os, on Plur. i, a 

ace. . . . on us, d 

dat. . . . u dbo (abo ?) 

gen. . . . i dn 

and these agree with the Gaulish endings of the a-declension, so far as they 

have been established.] 

64 EbeVs Celtic Studies. 

neuter, imned (tribulatio), pi. imnetha imneda. We recognize 
here distinctly the a-stem balla, fera instead of fir a, imnetha 
instead of imnitha ; fira harmonizes in a remarkably beautiful 
manner with the Gothic and Latin stem vira (for vair indicates 
a previous short i) in opposition to the Sanskrit vira. The 
feminine a-stems lead back to : — - 

Sing. Nom. -a 

Ace. -an . 

Gen. -e(s) or -(a)s 

Dat. -i or e 

Plur. -as 


Examples: nem (heaven), nem, nime, nim (stem nimd, hence 
the nom. nim is still found singly) ; delb (effigies), delb, delbe, 
deilb, plur. delbce (instead of delba), gen. delb, dat. delbaib, with 
primitive e, therefore it is in the dative not dilb, but deilb. The 
masculine stems, according to III., exhibit, in the immediately 
preceding stage approximately the following forms : — 

Sing. Nom. (-is or -us) 

Ace. (-in or -wi) 

Gen. «(s) or 6(s) 

Dat. u ? or -i ? 

Plur. -d(s), -e(s), i(s) 
-u -i 

-ibis (-abis ?) 

Examples: nom. denmid (doer) instead of denmadis, gen. 
denmada; nom. bith (world), dat. biuth instead of bithu; gnim 
(action) ace. plur. gnimu; aitribthid (possessor), gen. aitrebthado, 
nom. ace. plur. aitribthidi. 

It is easily seen that the forms which are attainable by imme- 
diate conclusion, do not admit, in any way, of a direct compa- 
rison with the primitive forms, as the Gothic, to a certain extent, 
do, but still require an intermediate stage to connect them. A 
baill ballu, or ballui, must necessarily have preceded balli, ballu, 
assumed from baill, baull, a nimd the nima, deduced from nem, 
a firm (oxfirun?) thefiru changed into firu, &firdn, the hypo- 
thetical firan in the gen. pi. In short, the oldest historical forms 
of the Irish, in regard to the conservation of the auslaut, stand, at 
most, and even scarcely, upon a level with the New High 
German, 42 as the simple comparison of the Irish and the German 
ball may show : — 



. Nom. ball, 

\ German, 

,, ball, 


„ baill, 


„ balle, 

Ace. ball, Gen. baill, Dat. baull. 

„ ball, ,, ball(e)s, „ ball(e). 

„ baullu, ,, ball, ,, ballaib. 

,, balle, „ balle, ,, ballen. 

We find that long vowels have disappeared in the auslaut 
often even with succeeding consonants ; equally so, short vowels, 
with succeeding s; only long vowels before s have preserved 

42 [xiv. Ebel would not now say this. See, infra, " On the so-called pros- 
thetic n", §. 12, p. 90.] 

On Declension in Irish. 65 

themselves in a shortened form: (forms such as cele (socius), con- 
sequently presuppose either a celias, celeas, with a fallen off end- 
syllable, or a celes with a shortening of the vowel before the 
fallen off s; we shall more correctly explain firu from firus 
than from firun, as we everywhere [except in the article aud 
teora n\ see that the long vowel in the genitive plural has dis- 
appeared along with the n). We could not, in view of such 
mutilation of the original endings, venture to think of anything 
like a satisfactory development of the case-endings, were it not 
that fortunately the above-mentioned law for the vocalism, and 
the changing of the consonants between the article and substan- 
tive, puts into our hands a test. 

The end-consonants, except m and r, have evidently all dis- 
appeared; m is changed, according to rule, into n, only traces of 
which have, still, been preserved ; 43 5 no longer occurs at the end ; 
t, which appears in its place in the Old Irish as int, and in the 
Modern Irish an t, shows us that it has only disappeared in the 
immediately preceding period, only after the dropping out of the 
short vowel. The Gaedhelic has, consequently, been harder than 
the Gothic, in so far that, besides s and r, it also suffered an n in 
its auslaut, probably derived, however, from m, not a primitive 
n. 44 Of these three consonants, s was the first which dropped off, 
for it does not appear in any declension or conjugation-ending; 
not even in the article, where, however, its former existence is 
betrayed by the t in the nom. int ant, and by the conservation 
of the original anlaut after the form inna na; the second that 
dropped off was the n derived from m, which is still visible at 
least in the article in the ace. inn, and in gen. plur. innan nan 
(besides here and there also, e.g. in teora ngutte, Zeuss. 310) ; 
r has preserved itself to the present day in the nominative athir 
athair (pater). 

The mutilations of the auslaut appear to have taken place in 
this wise ; in the first place the short vowels in the auslaut and 
before consonants were dropped, the long ones in the auslaut 
shortened, then (or also contemporaneously, a supposition to 
which the Lat. -um, instead of -urn, would lead us) the long 
vowels before n were shortened, hereupon s dropped, finally the 
long vowel was again shortened, and the short vowel together 
with n dropped. From the primitive Gaedhelic to the Gaedhelic 
of the oldest monuments, we would have, consequently, to pre- 
suppose three or four periods, which may be represented by an 
example, somewhat in the following manner: — 

43 [xiv. See the last mentioned paper.] 

44 [See on this passage the author's paper referred to in the last two notes ] 

66 EbeVs Celtic Studies. 

Primitive period. Pre-historic period. Historic period. 

Sing. Nom. . . . Dallas, balls, ball. 

Ace ballan, balln, ball. 

Gen balli, balli, baill. 

Dat ballui, ballu, baull. 


Plur. Nom. . . . balli, balli, baill. 

Ace ballus, ballu, baullu. 

Gen ballan, ballan, ball. 

Dat ballabis, ball(a)bis, ball(a)ib. 

Still later weakenings of the auslaut sometimes occur, as the 
Old Gaedhelic shows in neut. aill from aile (similar to the Old 
Latin alid) ; the Old Kymric especially distinguishes itself from 
the Gaedhelic by greater weakenings, e. g. as all (alius) and oil 
(omnis), instead of the Gaedhelic aile anduile. The adjective in the 
Welsh exhibits an interesting difference, inasmuch as here the 
change of i and u into e and o first takes place in the feminine, 
hence a fern, gwen, cron is opposed to the mas. gwyn (albus) crwn 
(rotundus). We may consequently presume that in the Welsh 
the fracture was only introduced when the shoit end-vowels were 
thrown off, consequently crunnas crunnd were already become 
crunn(s) crunna, whilst, in the Gaedhelic, the falling off only fol- 
lowed the introduction of the fracture. 

§. 4. Declension of consonantal stems. 

Now only are we in a position to attempt an explanation of 
the endings ; but, in consequence of the extremely difficult i- and 
w-stems, we shall begin with the declension of the consonantal 
stems. We find in Zeuss five classes (not exactly in the most 
convenient order), of which I. and II. contain n-stems, III. and 
V. /"-stems, IV. ^-sterns ; 45 of these d appears to have arisen out of t. 
The inflexion is most regular in the masculine-feminine w-stems 
(IL), and in the masculine J-stems (IV.). Both subdivide them- 
selves according to the vowel of the genitive into two divisions, 
in which we recognize, according to the phonetic laws of the Irish, 
stems with a and with i; those in -man may be compared with 
the Sanskrit -man, -iman, -van, and with the Greek -fiov (compare 
brithem judge and rjjEiuwv) ; those in -tin or -sin are, in a similar 
way, as in the Umbrian and Oscan, shortened from -tian, which 
again appears in the nom. -tiu, and consequently express the Lat. 
-tio, -tionis, with which they also agree in gender ; the infinitive use 
of these abstracts (comp. Zeuss, 462) explains the infinitives in 

45 [xv. Zeuss' series V. contains c-stems (in some instances z'-stems, which, in 
the oblique cases, go over to the c-declension), and under his fourth series he has 
put d-stems, ^-sterns, and ant-stems. Among his irregular nouns he gives ri, 
gen. rig, the sole example of an Irish ^-stem. Mi: (a month) gen. mis, is a 
ns-stem. So were the comparatives in iu, Sanskrit iydhs, though undeclined in 
the oldest Irish.] 

On Declension in Irish. 67 

-tinn, -sinn of the present language, which consequently are not 
at all directly connected with those in -t and -dh; probably a si- 
milar contraction of the stem lies at the basis of those in -id, be- 
cause in the nominative along with ogi (hospes),yiZi (poeta), tene 
(ignis), the fuller form coimdiu (dominus) shows itself. Analysis 
yields the common endings : — 

Sing. Norn, (long vowel) . . . Plur. -is 

Ac. -in {-en) . . . -as 

Gen. -as .... -an (-an) 

Dat. -i . . . -abis 

Which explain themselves without difficulty. The length in 
the accusative plural is remarkable ; it is proved by anmana (ani- 
mas), Jileda (poetas). As a change into the vowel-declension (like 
in the Latin -es,-eis, -is) in consequence of the a, in opposition to 
the -u or -i, which alone occurs in masc. vocalic stems, is not to be 
thought of, this -a must be either an inorganic lengthening, or -as 
has been produced from -ans, which has been already surmised 
to be the original ending of the accusative plural (Zeitschrift 
f. v. Sprachforschung I. 291, V. 63); the latter is probably the 
true explanation. Among the other endings, -as is remarkable 
by the peculiar tincture of the Gaedhelic vocalismus. For while 
the Greek, Latin, and Gothic agree in the weakening of the a 
in the genitive -oc, -us, -is, -is, in contrast with this in the Gothic 
even the nom. plur. -as remains pure, the Gaedhelic, on the other 
hand, in direct antithesis to the Gothic, has retained the genitive 
pure, — hence menman, noiden, druad, coimded, instead of men- 
manas, noidinas, druadas, coimdidas, and has weakened the nom. 
plur. to -is (or -es like Greek -tc?) consequently forming anmin, 
aisndisin, druid, filid. The accusative singular with its -in or 
(-en) may be compared with the Lat. -em, — in the Zend, even 
with a-stems, em, — hence menmain (for which also menmuin and 
menmin), airitin, torbataid or -tid, coimdid. The genitive plural 
has, of course, first shortened its -an to an, and then dropped it ; 
the dative singular may, probably, be referred as in the Greek 
and Gothic to the original locative. By the dropping off of the 
endings and the influence of the end-vowels, the gen. sing, and 
plur. on the one side, and the ace. and dat. sing, and nom. plur. 
on the other, of necessity became alike in sound. The dat. plur. 
took up a copulative vowel, as in the Latin and Gothic, an a, 
which by the influence of the dropped i has become ai or i; 
before this -aib, -ib, syncope frequently occurred as before the -a 
of the accusative plural, e. g. in traigthib (pedibus), always as it 
appears in the feminines in -tiu, the i of which, however, has 
acted upon the succeeding vowel ; hence dat. -tnib, ace. -tnea or 
tne. Zeuss' supposition of an accusative plural *druida, for which 

68 EbeVs Celtic Studies. 

we might expect *druada, appears to be erroneous. 46 We meet 
with various forms in the nom. sing, of em-stems, e. g.: masc. 
menme (mens), masc. brithem (judex) fern, anim (anima), fern. 
talam (terra) ; of the feminine m-stems passing into iu, sometimes 
weakened into -u; of the masc. ad-stems as a rule weakened to 
-u, and in tenge (lingua) to e; of -id generally -i, also, however, 
-iu in coimdiu (Dominus), -u in dinu (agna), and the adjective 
bibdu (guilty), -e in tene (ignis), gen. tened, stem tenid (instead 
of tanid as the Kymric tan shows); no ending in traig (pes). 
The form druith (druida), 47 from the stem druad, appears to 
depend upon the same transition into the i- declension as Lat. 
cards, juvenis, from the stem can, juven; for druith points back 
to *druadis. According to the analogy of the Sanskrit, the em- 
stems should have formed the nom. -a, which first was weakened 
to a, then fell off; brithem, anim, are, consequently, forms per- 
fectly in accordance with rule. The preservation of the vowel 
in menme, weakened, however, to e, appears to have been caused 
by the double consonants (as, perhaps, also in the gen. pi. athre, 
from athir, see further on). The -iu of the m-stems has arisen 
from the primitive -id (by passing through -ia or iu; the Lat. 
-io, Umbrian -iu speaks in favour of the latter), the u having 
been retained probably by means of the preceding vowel as in 
the dative celiu, as opposed to baull. The d- or ^-sterns pro- 
bably took originally, as in the Lat. and Greek, an s, lengthened 
the vowel before it as compensation for the t, and retained the 
shortened vowel after the dropping off of the s; e. g. *domnats 
(domnds) *domnus, *do?n?iu, domnu (profunditas). Or -ad was 
originally long, as shortening often takes place in the Gaedhelic, 
for example, in the adjectives in *acA = Kymric auc, awe (i. e. 
dc) ? In coimdid, together with coimdiu, shortening of the base 
of the stem may be assumed as the Welsh masc. in -iat (-iad, pi. 
-ieid), given by Zeuss (p. 806) come very near. Guiliat (qui 
videt) especially appears nearly to correspond to the Gaedhelic 
filid,^ the nom.Jlli would, consequently, be contracted from Jiliu, 
for which the dative duini together with duiniu affords an ana- 
logy. 49 Traig shows itself to be a Astern by Welsh troet, pi. 

46 [xvi. Druide is the ace. pi. in the Liber Hymnorum. This may perhaps 
have arisen, by progressive umlaut, from *druadi, if drui (like brathair) have 
passed over to the i-declension. The ace. pi. brdithre occurs in the epilogue to 
the Felire (609).] 

47 [xvii. Ebel has here been misled by Zeuss : druith. is the nom. dual, not the 
nom. singular, which must have been drui Q=.*drua(d)-s).'] 

48 [See " Note on a-, i-, d-, t- and nt- stems", §. 9, p. 83.] 

49 Zeuss, 755, considers the o?as primitive, and compares the Kymric -ed, -id, p. 
803 ; but, in my opinion, the masculine in -id ought rather to be compared with 
the Gaedhelic in -id, -aid, gen. -ada, and the Kymric -(/(now -dd) ; although 

On Declension in Irish. 69 

traet; Cornish troys, pi. troyes, treys; Armoric troad, pi. treid; 
but the nom. sing, traig and accus. plur. traigid are difficult to 
explain : the best way is, perhaps, by the assumption of a neuter 
(Zeuss, 274), by which the want of the ending would be jus- 
tified ; but the i in traigid is remarkable : we should have ex- 
pected *traigidd, *traigeda, traiged. Other deviations will be 
treated of hereafter; as regards cil (canis), whereof only the 
comp. banchu (bitch), and the derivative conde (caninus), occur 
in Zeuss, we may ascribe to the Old Irish the forms : ace. cuin, 
gen. con, dat. cuin; plur. nom. cuin, ac. cona, gen. con, dat. 
conaib.™ The neutral w-steins (I.) all derived with the suffix 
-man deviate from the expected form : — 

Sing. Nom. and Ace. -m . . Plur. -man (from -mand, manci) 

Gen. *-man . . -man 

Dat. *-main . . -manaib 

Putting aside slight fluctuations between a and e (e.g. nom. plur. 
ingramman, gen. ingremmen) in the gen. and dat. sing., the 
dative exhibits an exceptional m instead of n: anmim, anmaim 
(nomini), which appears to have arisen from assimilation ; the gen. 
anma, anmae, anme, has dropped the n. The remaining forms 
are made in a perfectly normal manner, but the nom. sing, ap- 
pears to have weakened the a of the original end -ma to i, be- 
fore it fell ofT, because of the continual occurrence of umlauts : 
ainm (nomen), beim (plaga), ingreim (persecutio), teidm (pestis), 
togairm (vocatio), senim (sonitus). 

The nouns of relationship in -thar (III.) contain the original 
a of the nom. sing, weakened to i, either by the influence of the 
liquids (Bopp, p. 1), or, as appears to me more probable, because 
the a weakened to a should have dropped out in the third period 
(as in balldn, ballan, ball); but this could not take place, in con- 
sequence of the unpronounceable double consonant (thr) thence 
resulting, and so at least the lightest vowel was chosen. The 
same reason caused, no doubt, the retention of the vowel in the 
gen, and dat. sing., the syncope of which was to be expected ac- 
cording to the analogy of other languages and of the plural cases 
(although a formation atharas, athars, athar, athari, athir, would 
not be impossible), and in the gen. plur. the retention of the end- 
ing-vowel in its weakened form e; bl at least, there is no reason to 
assume for the Old Irish a transition into the z-declension, which 

ancient, it is not primitive (compare Lat. lapid, Greek e\7rid, KopvO, Zeitschr. 
f. v. Sp. iv., 325, 332). 

50 [xviii. Kather thus : ace. coin h, gen. con, dat. coin ; plur. nom. coin, ace. 
cona, gen. con h, dat. cunaib.] 

51 [xix. This gen. plur. in e only occurs in athre, brdithre, and is certainly due 
to a passage over to the /-declension, Mdthair forms its gen. plur. regularly — 
thus : mdthar h.~] 

70 EbeVs Celtic Studies. 

to be sure would easily explain the form athre, but which even 
the Latin patrum spurned. In the dative plural, a, and not i, is 
also used as a copulative vowel, as athraib shows, 52 and if braith- 
rib occurs beside it, we must either view it as an invasion of 
the secondary i, or an indication of the early introduction into 
Irish of orthographical confusion. The nom. plur. is not sup- 
ported by evidence ; we cannot put it down otherwise than as 
athir, as Zeuss does. On the other hand, there is no evidence 
to entitle us to assume with Zeuss an ending -u for the masc, as 
we have no where detected, except in the nom. druith, a transi- 
tion into the vocalic declension. We accordingly assume the 
following genetic development : — 

Primitive period 

Pre-historic period. 

Historic period. 

Sing. Nom 




Ace. . . 




Gen. . . 




Dat. . . 




Plur. Nom. . . 



* athir 

Ace. . . 



* athra 

Gen. . . 




Dat. . . 




The addition of a determinative suffix already shows itself in the 
Old Irish in some r-stems (V.) ; in the Modern Irish its action has 
been felt over a much wider circuit, and has even penetrated the 
nouns of relationship. 53 Unfortunately, too few forms of this class 
have been preserved to us to give a complete idea of the declen- 
sion, nevertheless we see from the existing ones of cathir (oppi- 
dum) : — 

Sing. . . . cathir, cathraig, cathrach, cathir. 
Plur. . . . cathraig. 

— at least so much clearly, that these words, to which nathir 
(natrix) likewise belongs, even when assuming this suffix, fol- 
lowed a consonantal declension. Bopp's conjecture, adopted by 
Kuhn also, in his review (observation 15), that this ch (g) repre- 
sents an original k, is now completely justified by the Irish pho- 
netic law, according to which the tenuis between vowels changes 
into the aspirata (fluctuating into media) ; but to his comparison 
of the Gothic brothrahans and the Sanskrit -aha may be added 

52 [xx. In Gaulish e was used as a copulative vowel, as is shown by mdtrebo 
(matribus), cited supra. Note 23, p. 56] 

53 [xxi. This " deter minatrre suffix" is a dream. The Old Irish nouns to 
which Ebel alludes (though c'-stems in the nom. sing.), have, like yvvrj, passed 
over to the c-declension in the oblique cases. There are, of course, c-stems in all 
cases. Thus tethra, gen. tethrach (a scald-crow), is the Greek rirpa^ gen. 
rsrpaKog. The gen., dat., and ace. pi. of cathair may be set down with certainty 
as cathrach h, cathrachaib, cathracha, respectively ; for huasalathrach (patriarch- 
arum) occurs in St. Patrick's hymn {Liber Hymnorum), and huasalathrachaib 
(patriarchis) in Zeuss, p. 827 (the nom. sing, is huasalathair, cf. Ang"-Sax. heah- 
fcedher), and coercha (sheep, ace. pi.) for cderacha, in St. Brogan's hymn, v. 33.] 

On Declension in Irish. 71 

the still more apt one of the Greek -k in yvvi) yvvaiKog, like the 
opposite employment of the c in Latin, senex, senectus, along 
with senis (compare the essay of Ourtius on individualizing suf- 
fixes in Zeit. f. v. Sp. Bd. iv.) The dative cathir, no doubt, 
likewise rests upon a similar mutilation, as is frequently found 
among the ?i-stems, and should not have been placed by Zeuss in 
the paradigm; the normal form would be cathrich or cathraich, 
in the plur. ace. cathracha, gen. cathrach, dat. cathrachaib may 
be expected. 

In its most ancient stage the Gaedhelic, consequently, harmon- 
izes with the classic languages by the conservation of the conso- 
nant declension of the t-, n-, and r-stems ; it even exceeds the 
Latin in the conservation of the purity of the nom. ace. and gen. 
plur. ; on the other hand it associates itself to the Gothic by the 
passage of the s-stems into the vocalic declension, which takes place 
as in the Slavonic languages in two ways : by an addition in dis, 
disa, contrasting with the Sanskrit ciyus; by a loss in nem (nima) 
in contrast to the Sanskrit nabhas, with a change of gender, as 
in the Slavonic tima, against the Sanskrit tamas. 

§. 5. Declension of masc. {and neut.) a- and iA-stems. 

According to what has been said above, the vocalic declension 
includes masculine and neutral a-, i-, and w-stems, feminine a- and 
i- (i-) stems ; feminine w-stems are wanting, as in the Lithuanian. 

We have already carried back the inflexions of the masculine 
a-stems to the oldest attainable Celtic forms. Most of them 
scarcely require an observation. The nom. sing, -as, -a, -an, in- 
stead of -am, gen. plur. -an instead of -dm, agree exactly with the 
Sanskrit ; the dative plural -obis presupposes a more ancient pho- 
netic condition than we find preserved either in the Sanskrit in- 
strumental -dis or in the dative -ebhyas, and which is easiest 
explained from the instrumental (primitive form -abhis), for 
the dative form -abhyas would have led (through -abias -abeas, 
or through -abis -abi, through -abe*s -abS) to -abe or -aibi. bl (The 
-ai in -aib is not a diphthong, but umlaut, as the secondary form 
-ib shows ; it is, consequently, not comparable with the Sanskrit 
-3 m -Sbhyas). The dat. sing, -ui (or u? undoubtedly formed out 
of -ui) and the ace. plur. -us agree with the Lithuanian and Slavo- 
nian, being in the former -ui and -us, and in the latter -u and -y; 
the gen. sing, and nom. plur. -i agree with the Latin (besides 
the dat., Latin -6 from ~oz = Oscan -ui). In the nominative plural 
the pronominal ending (Sanskrit-e = primitive -ai, Lithuanian -ai, 
Gothic -ai, Greek oi, Latin i, older form -ei, Slavonian -i), has, 

51 [xxii. See note 23, p. 56.] 

72 EbeVs Celtic Studies. 

consequently, penetrated into the substantive declension in the 
Celtic also, as it does every where except in the Sanskrit, Gothic, 
Umbrian, and Oscan, and indfir (pronounced indir) from innifiri 
corresponds exactly with Mi viri; this i has, consequently, been 
formed out of -ai or -ex. On the other hand, in the genitive singular, 
the most difficult form, the -i corresponds to the Latin -z, which, as 
is well known, is written not -ei, but -i in Lucilius, and in the Sen. 
Cons, de Baca, an important circumstance for the correct explana- 
tion of the Latin form ; as for the rest, the explanation is easier in 
the Irish than in the Latin. Of the primitive ending = Sanskrit 
asya, not only y, which has everywhere fallen away, but also 
a vowel-flanked s must have disappeared in the Irish (Zeuss, 60, 
63); thus arose -ii (as in z'£A = Kymric, iot, /cc = Kymric iacc) 
which of course coalesced immediately into i; it only remains 
doubtful whether this -a also belongs to the Kymric or exclu- 
sively to the Gaedhelic. 52 The agreement of both forms with the 
Latin is, no doubt, the chief reason why the words borrowed 
from the Latin have mostly preserved, in so strikingly faithful a 
manner, the declension-type, and that transitions into this declen- 
sion have only taken place from the third Latin one ; — a change 
which the gen. -is induced, as, for example : socrdit, in conse- 
quence of socratis (even in the nom. preceptoir, plur. preceptor i, 
in consequence oft. preceptor is), not the reverse, except where it 
was necessary to join a word to a known ending, as in peccad 
masc, gen. pectha pecilio from peccatum, in consequence of the 
many words in -ad having similar meaning. The words in -e, 
sometimes written -a, and ya- (ia- and aia-) stems form a subdi- 
vision of the a-stems ; in them either -i before -a was changed into 
-e, or -ia was contracted into -3, -ii into -i, — these long vowels 
being naturally shortened in the auslaut ; all forms admit of being 
explained in both these ways in the most perfectly satisfactory 
manner. The -u in the dat. sing, remained here in the combina- 
tion -iu in the auslaut, for which, however, -u and -i also occur; 
in the dat. plur. a slight shortening took place, as iib did not give 
-ib, but -ib. bZ 

The neuters exhibit a curious anomaly, inasmuch as the prim- 
itive -a of the nom. and ace. plur., shortened to -a in the second 
period, should have dropped off in the third ; if we connect with -a 
of this case an analogous singular phenomenon, namely, that the 
inna, 7ia, of the article, as in the feminine, does not affect the suc- 

52 [xxiii. In the Old Irish, as in the Latin, the gen. sing, of masc. and neut. a- 
stems was originally the locative sing., and has nothing whatever to do with 
asya. Ebel is now inclined to admit this. See, infra, On the Position of the 
Celtic, §. 11, p. 125. 

53 Zeuss erroneously remarks, page 248 : quae -ib dativi non inficiens ex -ab de- 
fecisse videtur. The observation would have been in place at p. 253. 

On Declension in Irish. 73 

ceeding consonants, we shall be able to assume, with great pro- 
bability, that in the Gaedhelic the disappearance of the neuter, 
which in the Kymric can be no longer detected, had even then 
already been prepared in the plural, by the invasion of the femi- 
nine form, for the inna of the article does not admit of being ex- 
plained otherwise than from innds. The Irish na cenSla (nationes) 
consequently admits of being compared with the Italian le arme 
instead of ilia arma. Even the accusative plural masculine 
inna, na, appears to rest upon an inorganic invasion of the femi- 
nine form, because the substantive forms lead us to expect rather 
*innu, *nu [conversely -iu, (-it) = Lat. eos, occurs suffixed to the 
prepositions, even as feminine] ; this form has also penetrated 
in the Modem Irish, from the accusative even into the nomi- 
native, so that a difference of genders is nowhere to be found 
in the plural. The -ia stems form the plnr. nom. regularly in -e, 
as in the singular. 

The adjectives mostly follow the rule of the substantives, 
only that the ?<2-stems readily shorten the ace. plur. mas. into -i, 
and the nom. plur. neuter often shows -i instead of the more 
normal -e. The --/, which the a-stems often exhibit in the 
neuter plural, is mere remarkable, and is hitherto inexplicable 
to me. 54 A stem sdinia, instead ofsdnia, may probably be assumed 
for sain (di versus), in consequence of the ai. This has main- 
tained itself in the form of the nom. plur. ; in the others it has 
shortened itself like aile into aill. But how are we to explain 
isli, dilsi, comaicsi? Of the pronominal a-stems, a form has, 
however, been preserved, in spite of the frightful ravages here 
occasioned by the phonetic laws, which sets aside the only reason 
which could probably be still put forward (except the accidental 
similarity with the stem-auslaut a in the Sanskrit) in favour of ex- 
plaining the gen. -a of the following class by the Sanskrit -asya. 
Of the stem a, there have been preserved : gen. sing. masc. and 
neut. d, with affection of the succeeding consonants, consequently 
primitively a vowel-ending stem ; gen. fern, a without affection, 
consequently for as; gen. pi. an, a, consequently produced from 
an instead of am. Bopp therefore believed himself able to explain 
the masc. a by asya, and the fern, a (instead of as) by asyds. But 
now di appears as the most ancient form of the gen. sing. masc. and 
neut. (in Zeuss, 334, 345), besides ae, e (evidently e) also (Zeuss 
347) ; consequently asya modified itself in the first instance into 
di, and from thence issued the Gaedhelic forms a and e like the 

54 [xxiv. Adjectival a-stems never exhibit i in the nom. pi. But (as was to 
be expected) this is done by adjectival i-stems, such as sain, isil, ddis, comacuis, 
whence sdini, isli, dilsi, comaicsi. The adjectival z-declension exists at the present 
day. See the paradigm (geanamhail), O'Donovan's Grammar, p. 112.] 

7 B 

74 EbeTs Celtic Studies. 

Kymric y, e. Thus even tills form, which in consequence of its 
shortness must have sounded fuller, differs very little from the 
usual genitive of the a-stems. The neuter of the article an, -which 
has weakened itself even to a, rests no doubt on a primitive 
form anat,™ which from the outset must have become ana, an, be- 
cause anan (instead of anam) must have always retained an n; 
the fundamental -at also explains the more violent shortening in 
the neut. aill, as compared with the masc. and fern, aile.™ 

§. 6. Declension of masc. i- and u-stems. 

The explanation of the case-endings is much more difficult in 
the following classes, where the separation of the masculine u- 
and i-, and the feminine a- and i-, stems, is already difficult. 

The i- and ?<-stems sound in the nom. and ace. sing, perfectly 
alike, for -is, -in, -i must drop off like -us, -un, -u; even the vowel 
of the stem does not always give us information, although denmid 
(factor), for example, proves itself by the genitive denmada to 
have been altered from denmad, muir (mare) announces itself by 
its ui as an i-stem ; we must, therefore, endeavour to ascertain 
the stem from other sources, as, for instance, in bith (mundus), 
from the Gaulish bitu; in fid (arbor), from the Gaulish vidu and 
the Saxon widu; in the verbals in -ad, from the analogy of the 
Latin in -tus, etc. The only case which shows the stem clearly, 
the accusative plural, 57 the -us and -is of which have changed into 
-u and -i, is imfortunately only very weakly represented, so that, 
in many cases, no certainty can be attained. In the dative 
singular -ui and -i are certainly to be assumed; these should 
become -u and -i, and leave behind umlaut, but most words 
take no umlaut (no doubt, in consequence of the primitive 
length of the stem-vowel). Among the whole of the examples 
in Zeuss, biuth alone shows umlaut, which he accordingly has 
placed in the paradigm. It would appear as if the endings -a, 
-o, -e established a difference in the genitive singular ; but this is 
by no means the case, as aithrebthado, from the nom. aithribthid 
(possessor), for example, shows a decided z-stem; we must look 
upon -o rather as an obscuring of the -a, e, exactly as -ea and 
-eo are the result of the subsequent action of a preceding sound, 
or of one which had preceded. The explanation apparently 

55 [xxv. More probably tbe neut. article an (a before a noun beginning with a 
tenuis) stands for sa-n — the n being the neut. ending, and the sa the well-known 
pronominal stem. The s appears in composition with non-aspirating prepositions.] 

56 [For confirmation of this hypothesis see, infra, "On the so-called pros- 
thetic n", §. 12, p. 90.] 

57 [xxyi. The nom. and ace. plur. (-i) and dat. plur. (-i6) of z-stems show 
the stem clearly enough. But Ebel here, as elsewhere, suffers from the incom- 
pleteness of Zeuss's collection of examples.] 

On Declension in Irish. lb 

nearest at hand, that -o is derived from -aus ( = Sanskrit -6s), is, 
consequently, to be rejected, and we are to assume either that 
-aus, as well as -ais, has become -a, or, to start from the funda- 
mental form, -avas and -ajas, which must likewise become -as, 
-a; as the dative cannot be explained from- avi, -aji, the first hy- 
pothesis is, probably, to be preferred. 58 According to the analogy 
of the consonantal declension (compare also Gothic -yus and -eis), 
a fundamental form -avis and -ajis is to be laid down for the 
nom. plur. ; -ais must arise from -avis, and this, on the dropping 
of the s, could be contracted to -a, -e, or -i; -ajis, in consequence 
of the preponderance of the i-sound, passed, as it appears, exclu- 
sively into -i, certainly at least in the masculine in -ati (nom. 
-id, gen. -ada) ; the auslauts were, as everywhere, subsequently 
shortened, so that, along with -ai, -ae, -a, -e, and -t, also occur, 
e.g.: gnimai, gnimae, gnirna, gnime, gnimi, from the stem gnhnu 
(action). The form rnogi, from the stem mugu, along with mogae, 
is interesting, as their common origin from mogai is betrayed by 
their o. The ending -e of the gen. plur. is remarkable ; it appears 
to announce itself in moge as a degeneration of moga ; on the 
other hand, it has produced umlaut in forcitlaide (praacepto- 
rum) ; either there existed formerly a difference here, as in the 
nominative plural, so that -avan contracted itself into -an, -ajan 
into -ian, -en, or, the umlaut in forcitlaide is inorganic, and -e 
is in both cases degeneration of -a, from -dn=-avdn and ajdn, 
which forms we take as a starting point according to the ana- 
logy of the Gothic -ive and -e instead of -iye. The dative plural 
shows a remarkable anomaly, the normal -ib of the z'-stem in- 
deed appears in it, but not the -ub or -uib to be expected in the 
w-stem, but, instead of it, -aib (compare aitrebthidib , mogaib) ; 
either interchange has here taken place between ui and ai, a cir- 
cumstance otherwise without example (ui for ai is frequent), or 
the generality of the ending -aib introduced it inorganically here 
also, in the same manner as in the Greek iroXtcri, tt^xzctl the t ap- 
pears to have penetrated by means of the false analogy of the other 
cases. The neuter plur. in the nom. and ace. rind (constellations) 
mind (insignia), fess (scita), appears, at first sight, to be altogether 
anomalous without an ending, which is the more striking as even 
the a-stems show an ending where one ought not to expect it ; if, 

58 [xxvii. Surely it is easier to assume that the z'-stems (with one or two ex- 
ceptions, such as tir, tire) passed over in the gen. sing, to the w-declension. 
Hence the -o (-a) = -6s, -aus. The fern* a-stems likewise, in the gen. sing. — 
with five exceptions (inna, oena, mnda, cacha, nacha) — have passed over to the 
i-declension, and consequently exhibit the ending e = es, of which the e was 
probably produced, by a very ancient contraction, from a-i (cf. Goth, anstais). 
Here, of course, as also in the Sanskrit and Lithuanian dves, awes, " ewe's", 
the stem- vowel has been gunated.] 


EbeVs Celtic Studies. 

however, we start from a fundamental form -vet, -ja, in which the v 
sm&j were dropped, a development -d, -a, may also be conceived 
(perhaps we should even take a — ava, aja for a starting point, 
with inorganic gunation, in which case rind would bear the same 
relation to gnima, as ra\ka does to Ta\hg). In spite of much ob- 
scurity in details, it is at least clear from the preceding, that the i- 
and w-stems by no means so fully coincided from their origin, as 
would appear from the representation of Zeuss. For the sake of 
greater clearness, we shall here also attempt to give an idea of the 
declension arranged according to the different periods, without 
the secondary forms however : — 


Primitive period. Prehistoric period. 

Historic period. 


. Sing. 

Nom. . 








Gen. . 

(bithavas) bithas? 



Dat. . 





Nom. . 

(bitkavis) bithais 




(bithuns) bithus 



Gen. . 

(bithavan) bithavan 


* betha 

Dat. . 


bithui bs 

* bithuib 



. . . 





. . . 

(fidva) fida 






Nom. . 








Gen. . 

(denmadajas) denmadas? 



Dat. . 





Nom. . 

(denmadajis) denmadis? 

denmidi ? 



(denmadins) denmadis 



Gen. . 

( denmada j an) denmadajan 


* denmada 

Dat. . 






• • • 




PL , 

(fissja) fissa 



According to this view, it is only the dative plural of the 
w-stem mogaib that appears to be distinctly inorganic ; the gen. 
plur. moge shows a weakening of the a into e, which we shall 
presently find again in the feminine. 

§. 7. Declension of fern, a- and i- stems. 

The feminine a and i-stems have suffered still greater confu- 
sion in their declension, so that the primitive stem can now only be 
recognized from the vocalization of the nom. sing, and by com- 
parison with other languages. 59 Thus the following show them- 

59 [xxviii. It is true that in the Old Irish the fern, a-stems have in the 
gen. (but see note 58), dat. and ace. sing, gone over to the ^-declension; and 
in the dat. this was the case in Gaulish, as we learn from Belesami (nom. Bele- 
sama) in the inscription of Vaison. But in the Old Irish the fern, j-stems are 
f with very few exceptions*) still clearly distinguishable from the fern, a-stems. 

* Oabdil and its compounds are declined in the plur. like a-stems, so idbairt, ej/ert. 

On Declension in Irish. 11 

selves by e and o to be a-stems: ess, iress (fides), nem (ccelum), 
toft (voluntas), breth (judicium), crocli (crux), ingen (filia), 
aimser (tempus), and tlie words in -em, sucb as moidem (laus), 
cretem (fides) ; by ia instead of 3 — grian (sol), briathar (ver- 
bum), bliadan (annus); by comparison — run (mysterium) == 
Gothic runa, fere (ira)z=.bpy{), the words in -acht and -echt, 
which presuppose a Sanskrit -akatd and -ikatd, and which are 
not consequently derived directly from the stem-substantive, but 
through a hypothetical adjective in -ach or -ech ( = Sanskrit -aha, 
-ika), as for example, deacht (divinitas), which is not obtained 
directly from dia, but through *deach (clivinus). We must con- 
sider as i-stems especially the verbal-nouns in -t, such as epert 
(locutio), tabart, tabairt (datio), and also iarjigid (inquisitio, 
quaestio) ; the secondary forms, as muing, £. = mung, m. (a mane), 
quoted by Pictet, (Op. cit. p. 123), appear to be 2-stems (whose 
nominative -i, -i, >, cannot be distinguished in its actual state 
from -is, >s, >). No certain distinctions can be at all recognized 
in the case-endings, and nothing can be based upon the secondary 
forms. The genitive singular shows, for instance, along with 
the dominant -e, also -a and -o; but if we would assign the -a to 
the «-stems, and the -e to the z-stems, we find our proposition con- 
tradicted by the circumstance that -e is the commonest ending, 
and appears just in those words the vowels of which point to -a, 
as in nime, irisse, ingine, and that -a occurs frequently in charac- 
teristic z-stems, as in eperta; if, on the other hand, we would 
assign -a to the ^-sterns, from the analogy of the masculine, and -e 
to the a-stems from the analogy of the Latin -ce, the feminine of 
the adjectives like cacha, nacha, (and even dena, along with aine), 
will remain unconsidered; consequently -a is clearly the oldest 
form in both classes, it weakened itself into -o and -e, even in the 
same words; e. g., duile and dulo, from dul (mundus, res, crea- 
tura), and the umlaut before e, in spite of its universality, is in- 
organic ; the fundamental forms -as and -ajas had also to follow 
the same course : -as, -a, -a, or if we prefer starting from -ais in- 

In addition to the circumstance that the a- stems in general have their gen. sing, 
in - e, whereas the i-stems make it in -o (a), the nom. and ace. pi. of fem. {-stems 
end in -i, but those of the a-stems in -a. Next, the gen pi. of fem. i-stems ends in 
ae, -e; that of fem. a-stems has no ending. Thus nime, dule, caille, rigne, infinite, 
bliadne, fochraice,Jochide, are the Old Irish genitives plur. respectively of nem, 
nim (heaven), dull (a thing), caill (a wood), rigain (a queen), infinit (an infinitive), 
bliadain (a year), (not bliadan as Ebel wrongly gives it) ; fochricc (a reward), 
fochaid (tribulation). Thirdly, the dat. pi. of fem. i-stems ends in -ib, that of 
a-stems in -aib (dirmib, Zeuss, p. 670, probably comes from * dirim : cf. Welsh rhij ).] 
60 In the Lord's Prayer, as given by O'Donovan, there is, however, bid do toil 
(thy will be done), which indicates an i-stem.* 

* [xxix. Toil here is the accusative sing., according to the regular Old Irish syntax 
(Zeuss, p. S94) : the nom. sing, is lol, which was anciently a fem. d-stem.] 

78 EbeVs Celtic Studies. 

stead of -ajas, we have -ais, -ai, -a. The i -stems could form the 
dat. sing, in 4, -i (or -aji, i, 4, which is less probable), the a- 
stems either in (-di), -4, e, or (-ai), 4, i-, as in the nominative 
plural of the masculine ; both of them consequently agree, as may 
be expected, in the umlaut. An -is, 4, -i might have been ex- 
pected in the nominative plural, as in the masculine, from the 
fundamental form -ajis; but an ais, -ai, -a, was equally possible; 
and if the examples give -a, -e, and 4, an -ai, 4, 4 is not impossible, 
even in the case of a-stems (compare Greek -at, Latin -ae) : con- 
sequently a separation of both classes, according to the ending, 
is neither a priori necessary, nor in the actual state possible (see 
the examples in Zeuss, 262, 263) ; although, no doubt, the as- 
sumption of a primitive difference between -a (from -as) and 4 
(from -ajis) would have much in its favour. What is most striking 
is, that no ending whatever is found, not only mpersin from persan 
(persona), which is treated in Modern Irish altogether as an «-stem 
(nom. pearsa), but also in aimsir; and only in the vowel is there 
an indication of 4. Zeuss considers the e and i as secondary forms, 
which have resulted from assimilation: litre, epistli, appear to 
speak in favour of this view, but not bliadni; for an a has been 
here dropped. The following hypothesis appears to me to offe] 
most advantages : the feminines in 4 formed like the masculines 
the nominative plural in 4 (see above), those in -a, contracted -di 
(as in the Greek and Latin), into e or /, which, in consequence of 
its genesis from -di, yielded somewhat more resistance to re- 
trenchment than the 4 of the masculine resulting from -ai, and 
which therefore maintained itself, in part, in the weakening -e, 4, 
and in part actually dropped off; but the form -a rests (as in 
Slav, -y, -e), on an interchange with the accusative, which already 
in some instances took place in the old language, but which has 
deformed the whole declension in the modern. This hypothesis 
is supported by the nominative plural of the z'd-stems, which never 
contain -e, but everywhere 4; a circumstance which points to an 
earlier 4 generated from 4e or 4i. The class -distinctions are com- 
pletely obliterated in the gen. plur. (without ending), dat. (-aib 
and 4b without distinction), and ace. plur., 61 which often termi- 
nates in ~a even in undoubted i-stems, e. g., idbarta (oblationes), 
seldom in 4, as duli (res), epistli (epistolas). 

If almost everywhere here, an invasion occurred of the most 
numerous a-stems, the reverse appears to have taken place in 
the accusative sing., which exhibits, almost without exception, 
umlaut or a primitive i; only delb (imaginem) and nem (caelum) 
point to an ending -an (an). Even if we were to assume that -an 

61 [See Note 59, p. 76.] 

On Declension in Irish. 79 

was changed, as in the Zend, into -en (in the consonantal declen- 
sion we were led to an accusative -in or -en), the cause why this 
degeneration did befall the primitive -an of the feminine rather 
than the -an of the masculine, would still remain unexplained. 
The m-stems partake of the above mentioned deformities in the 
accusative singular, which terminates in -i instead of -e, and in 
the accusative plural, which likewise ends in -i, on the other 
hand the gen. sing, -e leads us back to the primitive -a of this 
case ; the nominative plural -i appears to be formed according to 
rule, except that all the end syllables are shortened. Accordingly, 
instead of the forms to be expected, — which are somewhat as 
follows : 

Sing. Nom. 

-a -a — 


>s > 


-an -an — 


In I 


-as -a -a 


-a -a 


-i -i > 


-* I 

Plur. Nom. 

-X -i )(?) 


-i -i 


-as -a -a 


-i -i 


-an -an — 


-an -a 


-dbis -aibs -aib 


-ibs -ib 

-we find the following actually occurring : 


>_ (-) 
> e (-a, o) 



-a (-1) 
-aib (-ib) 

in which d represents the after-action of the retrenched i. The 
same degeneration of the original forms occurs, as may be ex- 
pected, in the Modern Irish, where an cholam (columba) fluctuates 
in the gen. sing, and nom. plur. between na colaime and colama, 
and even in the dat. sing, between do'n cholam and cholaime; it is 
still further increased by the circumstance that the genitive has 
also frequently thrown off the inflexion vowel, e. g. na hoigh from 
an oigh (virgo). In general, however, the a-stems appear to have 
assumed the ending -e; the z-stems on the other hand -a, e.g.: slat 
(rod), gen. sing, and nom. plur. slaite; sgiath (wings), gen. sgeithe; 
neamh (heaven), gen. neimhe; hutfeoil (flesh), has however, gen. 
sing, and nom. plur. feola; and oigh, although in the gen. sing., it 
has hoigh, in the plural it is na hogha. The fluctuation has even 
passed over to the masculine, for iasg (fish) forms gen. disc, plur. 
disc or iasca; and sruih (scholar), in both cases smith or srotha. 
Already in the Old Irish, the vocative has been replaced through* 
out in the plural by the accusative ; in the singular there are only 
some forms of the a- and a-stems preserved, e.g. fir horn, fire, as 
in other languages ; duini from duinie; and among consonantal 
stems the single one ath(a)ir in the Lord's prayer. We have 
already found in the Old Irish beginnings of a permutation of the 

80 Ebel's Celtic Studies. 

accusative and nominative. The consonantal n- and ^-sterns suffer 
likewise a peculiar mutilation in the Old Irish. The secondary 
forms of anim (anima) ; gen. a?ime, dat. and ace. anim, admit of 
being explained from a vocalic base : not so the anomaly, which 
not unfrequently occurs, that the nominative directly supplants 
the dative and accusative. Examples: do foditiu (ad tolera- 
tionem), do aurlatu (ad obedientiam), ace. aurlatu (obedientia) ; 
compare also Pictet's observations (Beitrage zur vergleichenden 
Sprachforschung, I. 82 sq.), where the reverse is likewise proved. 
The circumstance that, in the Modern Irish, there is mostly (ex- 
cept in the anlaut) no difference to be found between the nomi- 
native and dative singular, agrees with the foregoing; it con- 
sequently appears that the accusative first was identified with 
the nominative, and then the dative. The language is, therefore, 
in a fair way to lose all its inflexions like the Kymric dialects, 
and first of all the genitive plural, which now is already mostly 
like the nom. sing. ; — properly speaking, only the gen. sing, and 
plur. and dat. plur. are yet retained : nay, even the latter has been 
already deprived of its ending in the article, in the same way as 
the adjectives have lost all their inflexions. The decision as to 
the origin of the modern forms of the consonantal stems is ren- 
dered more difficult by this phenomenon. Only few still corre- 
spond to the old form, thus breitheamli (judex), gen. breitheamhan, 
nom. plur. breitheamhuin, with brithem, gen. britheman, nom. 
plur. brithemain. DaileamJi (butler), for example, deviates al- 
ready in the gen. daileamhidn, from ddlem (caupo), gen. ddleman. 
The majority have affixed -e or -a either in the nom. plur. or in 
both cases, and it is difficult to decide whether we are to look upon 
this as a simple transition into the vocalic declension (as in New 
High German brunnen, instead of brunn), or whether the nom. 
in -a is not really an accusative ; perhaps the accusative form first 
passed into the nominative, and then the genitive singular fol- 
lowed the analogy of the nominative plural now appearing vo- 
calic. A striking example of this mixture of forms is afforded 
by cu (cards) ; gen. con (perfectly normal), or cuin (a- stem) ; 
dat. coin (normal) ; nom. plur. cona (accusative form), or con (spu- 
rious formation), or coin (normal) ; gen. cu (mutilated), or con 
(normal) ; dat. conaibh. The nominative plural athara from athair 
(father), has assumed the accusative form, and thereby got the ex- 
ternal appearance of a vocalic stem, an example in which it was 
followed by the gen. sing, athara (in use besides the primitive 
athar); side by side with them forms with -ach have been intro- 
duced; e. g.: aitlireach (as in Old Irish cathir). 62 The applica- 

62 [xxx. Aitlireach is simply due to a passage over to the c- declension. So 

On Declension in Irish. 81 

tion of the suffix -adh (compare denmid, denmada, or tenga, ten- 
gad), as an inflexion-copulative, is new; e. g., in the plural bo- 
gadha (for bogha, bows), considered also by Pictet (Op. cit. 128) 
to be a new formation ; but, perhaps, it may help us to an expla- 
nation of the Kymric plural forms. 

§. 8. The distinction of the plural in Kymric. 

The Kymric, on which we must in conclusion cast a glance, 
has preserved nothing more of its whole inflexions, even in the 
oldest documents, than the distinction of the plural, but this it 
employs very arbitrarily: compare trimeib (tres filii) with mei- 
bion, meibon, and tyreu (turres) with tyroed. Obviously, as in the 
New High German, this is of three kinds : either the old plural 
form remains, consequently true inflexions, as brilder, gdste, 
fische, from the Gothic brothrjus, gasteis, JisMs; or the ending 
of the stem, dropped in the singular, behind which the gramma- 
tical ending has disappeared, as in mannen, where the -an of the 
Gothic manna (stem mannart), which has vanished in the singular, 
has been preserved, while the proper ending, the s of mannans, 
has been dropped; or a suffix (determinative), wholly foreign to 
the stem, like the German -er in eier, to which true inflexion- 
endings were, at an earlier period (Anglo-Saxon dgru), attached, 
but wlrich, after their loss (as in the Old High German nom. 
eigir), exactly occupies the place of the ending, like German 
lander instead of lande, except in the dative plural. 

To the first kind belong : 1, the Kymric plurals without end- 
ings, and with umlaut, such as Welsh llygeit = Cornish legeit 
(oculi) ; Welsh seint=z Aimoric sent (sancti) ; Welsh chwaer (soro- 
res), from chwior; tract = Cornish treys, Armoric treid (pedes), 
from troet, Cornish troys, Armoric troad, — or without umlaut, as 
tridyn (tres homines), teir morwyn (tres puellae). 'All these 
forms have lost an -i, probably a primitive 4 or -is (-is?), 
and consequently may be compared to the Gaedhelic forms 
such as maicc (filii), to which the Welsh meib, or traigid, the 
Kymric traet, treys, treid correspond; for instance, the mascu- 
line verbals in -iat, -iad, pi. -ieid, such as guiliat, are parallel 
to the Gaedhelic in-z, pi. -id (filid) (see above). 2. The plurals 
in i, such as meini (lapides), from maen, Corn, esely (mem- 
bra) = Armoric ysily, from esel, appear to correspond to the 
Gaedhelic -i (in ia- and feminine stems) ; but interchanges occur, 
however, such as Cornish meyn, Armoric mem, alongside of 
Welsh meini, and this even in the same dialect, e. g. : Cornish tell, 
and also tylly (foramina), from tol, which do not allow a strict 

in Early Middle Irish we have mainistir (from monasterium),.makh\g its gen. sing. 
?nanestrech. Zeuss, xxviii. so altdir, from alture, gen. altdrach.'] 

82 EbeVs Celtic Studies. 

separation to be effected. As further instances may also be ad- 
duced llestri, Cornish, and Armoric, listri, "which represent 
Gaedhelic *lestir, while on the other hand dyn is the Gaedhelic 
doini. 3. Finally, the plurals in -au and -iau with their different 
formations (Zeuss, 290, 122), also belong originally to this 
category; e. g. tyreu (turres), Cornish detliyow = Armoric diziou 
(dies) ; -au appears to have belonged originally to the w-stems, 
the verbals in -at (-iat), -ad, pi. -adau also correspond to the 
Gaedhelic abstracts (infinitive) in -ad, -ud, which take -a in 
plural, so that -au may be very well explained from the Sanskrit 
-avas. Pictet's {Op. cit., p. 135) comparison with the Sanskrit 
-as, which changes into -6 before sonants, although adopted by 
Bopp and Kuhn also, is certainly erroneous. But afterwards 
confusion came in here likewise, so that we see -au exactly like 
the Slavonian -ov and the Greek -ev and other determinatives 
applied to other stems also, and hence even arose -iau. Besides, 
all three suffixes occur in both genders, so that perhaps the -i of 
the feminine may confirm the above assumed Gaedhelic funda- 
mental form of the nominative plural. 

The second kind embraces especially n-stems, such as the ap- 
parently anomalous hi (canis), the plural of which is in Welsh, 
cdn, cwn, Cornish ken, and which corresponds exactly with the 
Gaedhelic cu, plur. cuin (the Gaedhelic u is the Kymric i); and 
z/c/i = ox, plur. y chain (ancient, ychen) = oxen ; — further, Welsh 
brawt, which has lost its final r, plur. brodyr, (Cornish brand and 
broder, while in the Armoric sing, breur, breer, the d has yielded, 
plur. breuder). 

Kuhn (p. 595) wished also to include under the third category 
the -an of gen. cluasan (the ears), but in this word it belongs un- 
doubtedly to the third, as cluas is evidently the old stem, which, 
in the beginning, was treated in the declension like dis. 

To the third kind belong the following : 1. Many plurals in -au, 
-iau, in which the ending is foreign to the word-stem proper, such 
as penneu (capita), stem pinna (or pinda) = Gaedhelic cinna, from 
which nom. cenn, dat. ciunn, or breicheu (brachia), stem toeich, 
instead of breclii ; 2, most words in -ion (or -on), e.g. — deneon, 
dynyon (homines), from the stem dini (instead of dinia, as the 
Gaedhelic duine shows), or meibion (filii), along with which appear 
likewise after numerals the forms meib, dyn, and all Welsh plural 
adjectives, e.g. meirwon, along vnthmeirw, from marw (mortuus) 
= Gaedhelic marb, plural mairb (moirb). The -n consequently 
takes exactly the same place here as in the German adjectives 
and many feminities. 3. The endings -et, -ot, -ieit, -eit, and -ed, yd, 
oed, which otherwise occur as derivatives, and in this respect have 
been already compared above with the Gaedhelic -ad, -id, likewise 

On Declension in Irish. 83 

join many stems as determinatives, in which respect they are 
parallel with the -ad, in Irish bogadha, already compared, if I am 
not mistaken, by Kuhn. (Both forms are related to one another, 
as x a 9 lT 1S to ^ 7rt ^ in the Greek.) Compare the following 
words in -t: merchet (flliae), from merch (is this identical with 
Lithuanian, merga? cf. p.), Cornish denys (homines), Armoric 
bretonet (Britanni) with those in -ed : Welsh, bydoed (mundi) 
from bytz=z Irish bith, Cornish eleth = Armoric aelez (angeh). 
On the other hand, the favourite suffix of the Gaedhelic -adh is 
not employed as a determinative in Kymric. 

In the representation of my results, I have altogether followed 
the same analytical way which I had gone in the investigation 
itself, in order to rendei the verification easier to the reader. 
Some points will require completion and correction. On the 
whole, I hope that the results obtained will be found correct. 

§. 9. Note on a-, /-, d-, r-, and nt- Stems. 

According to a communication of Mr. Stokes, that has reached 
me through Professor Kuhn, 63 the a-stems show in the Old Off am 
inscriptions not only the gen. in i — MAQVI 64 (a form which ex- 
plains by its qv not only the Kymric map, but also the Gaed- 
helic mace without aspiration), — but also the nominative in -as 
(CORPIMAQVAS— Cormac). This highly interesting form 
may accordingly be placed by the side of /uapicav, Pausanias, x. 19, 
11, in which we are now justified in recognizing the true Gaulish 
accusative of marcas* (ngen. marc^ w. 3, march, plur. meircli). 
The Ogam secondary forms in -os, show us at what a remote 
period the obscuration of the a to o was already common. I 
would not, with Stokes, 65 deduce the length of the dat. plur. from 
the single form sceldib, as even feminine a-stems fluctuate between 
-ab, -ib, aib, which indicates a short vowel; and the id-stems 
invariably show -ib, instead of the -ib to be expected. 

That the neutral aill rests on a vocalic fundamental form, the t 
or d being dropped (like Greek a\Xo), as was already suspected (p. 
90), is confirmed by the mortification of the s in alaill sain, Z. 364. 

According to an observation kindly communicated to me, 

Mr. Stokes now recognizes in Zeuss' Ordo Posterior Ser. 4., 

three kinds of stems, in -d, -t, and -nt. The latter, to which dime, 

fiadu, car a, ndma (ndmae), belong, correspond accurately with 

the participles in -ant, 66 as, for instance, car a (from cairim, amo), 

fiadu {^vedant, — Stokes) ; dinu appears to be connected with the 

63 [Published in the Beitrage z, v. Sp. i. 448.] 

64 [Given in Mr. Stokes' paper, "Bemerkungen iiber die irischen declina- 
tionen"— Beitr. z. v. Sp. i. 333.] 

65 [Idem, 336.] 66 Also, Stokes' view, Beitr. i. 457. 

84 EbeVs Celtic Studies. 

Sanskrit root dhe (" suckling") ; cara and ndma likewise occur in 
the nom. in Zeuss, who has mistaken the true relation, and led 
rne astray: imcara fa aescare (sive amicus, sive inimicus), 674, 
831, and bannamae (inimica), together with the ace. bannamit 
(hostem), 820, the ace. carit, 1055, 1062, escarit, 1056. These 
stems appear to be of the common gender like the Latin participles. 
On the other hand, the -it in nebcongabthetit stands no doubt er- 
roneously for -itk (as generally in all abstracts). That traig is a 
neuter appears to be confirmed by traig cethargarait, 1018 (Gl. 
proceleusmaticum, consequently an ace.) ; it looks like a participle 
( = rp£Yov), but inflects the dat. plur. traigthib, ace. plur. traigid; 
traigthech (pedes, pedester), and traichtechdae, instead of triag- 
thechdae (pedester), are derivations ; the neuters have, therefore, 
perhaps thrown out the n, and taken a weak form (traigthib = 
tragitdbis). The Kymric troet, plur. traet, appears to rest on stem- 
extension, — compare Welsh, 2. cilid, 3. cilyd, with Gaedhelic 
cele; at least, a Kymric car, tan, stands parallel with the Gaedhelic 
cara, tene, so that we have to recognize in the Kymric forms 
rather the nominative, than, as in the Romance languages, the 
accusative (see further on). The comparison made in the' article 
on declension (page 68) between the Kymric guiliat and the 
Gaedhelic filed falls to the ground with the explanation of Zeuss ; 
see the corrections to pages 149 and 806, at the end of the 
Grammatica Celtica. 

I cannot as yet make up my mind to give up my former view 
respecting the feminines in the Ordo Prior, Ser. 5 of Zeuss, 
namely, that an almost complete fusion of the i- and a-stems took 
place, and that only few relics of a stricter separation of the 
forms have been preserved. Along with the ace. plur. in -i, to 
which suliZ. 339, likewise belongs, there occur, however, forms 
with -a from undoubted ^-sterns, as gabdla ; along with the 
dative in -aib, forms occur in ib from a-stems, as airmib from 
dram, slebib from sliab; so that nimib also does not prove a stem 
*nami (the nom. nim along with nem, ace. nem, the adjective 
nemde=*nimatya seems to point to *nimd, as also the Kymric 
nef, which perfectly corresponds to the feminine of the adjective in 
the Welsh, while i, u, disappear without umlaut in the Kymric ; 
further, that nem- never occurs before the endings with e, t, but 
always nim-; the gen. plur. nime is however remarkable). But I 
cannot adopt Mr. Stokes 1 view about the gen. sing, in -e, -a ; for, 
in the first place we should not start from Sanskrit -es, but from 
the fundamental form -ais (or ay as?), out of which -a (o), and -e 
could be developed in the masculine stems ; but -yds is a special 
Sanskrit form, which does not again occur in any European 
language (for that ttoXswq is not to be explained from it, but 

On Declension in Irish. 85 

from *7roXf?/oc, is proved by the Homeric ttoXvoq, the unjustly 
attacked masc. fiavrvog, and the neuter acrrswg, which, although 
questioned, is a well-attested form with the Tragic Poets); se- 
condly, because umlaut is as little known before a (o) among i- 
stems as a-stems: compare jlatha , flatho , or even focheda, fochodo ; 
a occurs even before -e in ergabale; we could not consequently 
lay down as a basis any such form as -yas, and must, as I believe, 
assume that the umlaut in both classes has only been introduced 
inorganically with the change of the a into e. &7 The analogy of 
the gen. plur., especially the invaluable nclndida™ appears even to 
speak in favour of our starting, both here and in the masculine of 
Ser. III., from -ajas (not from -ais). 

As regards the z-stems, it appears to me more and more pro- 
bable, that they have almost throughout passed, as in the Greek, 
into the za-class (7r6Tvia=pat?ii, etc.) 

I have found the umlaut in the dative of the w-stems, in 
immognom, Z. 984. 

§. 10. On the Celtic Dual. 

Agreeably to the wish of Mr. Stokes, I here give my views 
about the Celtic Dual. It appears to me that the answering of 
two preliminary questions is in the first place needful: 1. has 
the Celtic a dual to show? 2. how much of it is preserved? 

As regards the first question, there can be no doubt that the 
declension of the numeral two presents us with true dual-forms ; 
for the nom. and ace. masc. da (as it stands written in all ex- 
amples, more correctly however da, compare ddu, Zeuss 369, 
and Welsh 1. 2. dou, 3. den, now dau) exactly represents the 
Sanskrit dvdu, Latin duo, Greek Svo for the older Suw (Sfw in 
SwdeKa), and the primitive vocalic ending is proved by eter da 
son. Z. 197. The nom. and ace. fern di = Welsh 2. did, 3. dwy, also 
agrees exactly with the Sanskrit dve, Slavonic dime, Lithuanian 
dm; the dative deib ndillib evidently points back, according to 
the correct observations of Stokes, to a *dvdbhim weakened 
from dvdbhydm (or rather *dvabhim, cf. Svotv instead of §v6(piv). 
We have consequently also to refer the genitive da to *dvaans = 
Skr. dvayos, at all events the aspiration in da charpat is erro- 
neous ; 69 the n in the nom. and ace. neut. is however difficult to ex- 
plain. But that dual-forms are likewise preserved in the declen- 
sion of substantives, is proved by the peculiarity of the Kymric 

67 [See notes 58, 59, pp. 75, 76.] 

68 [xxxrv. Dula is, unfortunately, only found in a Middle Irish MS. : in Old 
Irish MSS. it is always either dale or duile.~] 

69 [It is possible that the aspiration after the genitive dual is correct, as this 
case ends only in Sanskrit in s, but in a vowel in Zend, Lithuanian, Slavonic] 

86 EbeVs Celtic Studies. 

dialects to put, after the numeral 2, the same forms as in the 
singular. The Welsh uab instead of mab in (W. 3) deu uab, — 
the Gaedhelic mace in da mace, is evidently as little a true sin- 
gular form as the Gaedh. fer after cet and mile is a true nom. 
sing. ; but the form of the nom. sing comes just as well where it 
distinguishes itself from the only conceivable genitive plural, as 
here, where the greatest similarity exists between the genitive 
plural and the nominative singular ; in deu uab = da mace a true 
dual has consequently been preserved (as the primitive form of 
substantives has generally been preserved in the Kymric after nu- 
merals, e.g. trimeib = Gaedh. trimaicc, that it is *tris maqvi, instead 
of the usual meibion), and the agreement of the nom. dual with 
the nom. sing, in most cases, caused by the Celtic phonetic laws, 
has led in other cases to an unwarranted extension of the singular 
form. The Celtic with its dual in the nom. of substantives stands 
therefore in an interesting contrast to the Teutonic languages, 
which had already lost the dual in the substantive in its earliest 
stage, but have preserved it in the Gothic verb. 

But the detection of the nom. dual leaves the second question 
still unanswered. Even in the Greek the genitive-locative is 
lost, and replaced by the form of the instrumental-dative-abla- 
tive ; duo and ambo in the Latin have not remained in undis- 
turbed possession of the accusative, indeed the nom. is replaced 
in the feminine by duae; nay even the Lithuanian, notwithstand- 
ing its close affinity to the Slavonian (the only European lan- 
guage which has completely preserved the dual in all forms), has 
undoubtedly lost the locative, and very probably replaced the 
genitive by the genitive plural (in spite of Bopp's opposite view, 
compare Gram. I. 2 Ed. 442; Schleicher, Lith. Gram. 171; — 
according to Schleicher Beitr'age, I. 115, s is not dropped in 
Lithuanian). It need not therefore at all surprise us, if all the 
dual cases have not been preserved in Irish, and the less so, as 
the Gaedhelic, like the Kymric adjective, always appears in the 
plural : Gaedh. da druith aegeptacdi, da ngruad corcra, da nainm 
cosmaili; W. (3) deu was ieueinc. In fact it may be proved, 
that even the substantives of the ordo prior (see Appendix I.) 
series 2 and 5, consequently a-stems and feminines, and all con- 
sonant stems {ordo posterior), have lost the genitive dual, and 
replaced it by the genitive plural. The primitive ending of 
this case -aus = Skr. -6s, could scarcely ever (if the phonetic 
laws laid down in Gram. Celt. I., 165 sqq. are correct; and that 
they are, the almost transparent clearness in which the greater 
part of the case-endings appear according to them, is a guaran- 
tee) so wholly disappear as that, in Old Irish at least, an -a as 
a contraction of -a or -au would not have remained ; but as we 

On Declension in Irish. 87 

find not alone from consonantal stems da arad, but also from 
a-stems da-tarb, da mace, da charpat (instead of carpat), da 
lethcend (no doubt more correctly lethchend, as a vowel (i) has 
dropped out in the composition, stem letldz=.\j2X. latus, Gr. 
rrXarog, letliclienn is ri/uiKpaipa), da carachtar, nay even from 
a-stems (indarann) without endings, we must look upon them as 
genitive plurals, which, as in the Lithuanian, have taken the 
place of the genitive dual. (To the preceding examples may be 
further added a cla'syl. 7a. 369, that is a da sillab, with a wrong 
mortification point ; I am in doubt about the stem of da og, whe- 
ther it is ogi or oga?) The form of the article in, also, which 
even as arising from innan is very strange, does not admit of 
being at all explained from innds (innaus). The dd in inddd 
aimserda is probably only a sign that dd should not be aspirated. 
In the same way we shall consequently have also to explain the 
forms of the m-stems — dagutae, inda gutae fern. ; z-stems — inda 
leitliesin (n. or m.) ; ^-sterns — inddd aimserda, da lino, which 
might admit of being explained perhaps otherwise also as real 
dual forms. It is evidently an accident that we should find just 
here a form in -o among i«-stems, while the genitive plural other- 
wise generally ends in -e, and only once in -a (Stokes, Beitr. I. 
346) ; and least of all should it have misled Zeuss to place even 
in series 1 and 2 the form of the genitive singular in the para- 
digma. The dative remains doubtful, as the whole of the forms 
may be explained as well from -bin as from -bis, and the Greek 
and Lithuanian have just preserved this case: indibmaigib, dib 
cetaib Z. 311, 313, deib hdillib, clondib dligedib remeperthib , dib 
rannaib, dib consonaib 194, indib nuarib deac, dimutaib, deib 
traigthib; I would, however, almost prefer, here also, the expla- 
nation as dative plural, because the Celtic has retained so very 
much less of the dual than the Greek and Lithuanian, no verbal 
and no adjectival forms. 

Of undoubted dual-forms we accordingly have only the nom. 
and ace. of substantives, and the whole of the cases of the numeral 
two. The masculine a-stems, with the exception of the one- 
syllabled dd, da, have thrown off the ending -a, shortened from 
-a, ( = Ved. Slav, -a, Gr. -w, Lat. o) or -au ( = Skr. -du), hence 
nom. da mace, da mod, da son, ace. indamer (?), inda articul, eter 
da son, Z. 197; the ar II. canoin (pro duobus canonicis) kindly 
communicated by Mr. Stokes, is consequently to be completed 
dd canoinech (more correctly chanoinech.). The neuters, in de- 
viating from the Sanskrit and Slavonian, connect themselves with 
the Greek and Latin, inasmuch as they likewise presuppose an 
ending -a, older -a (or -au) — da ngruad corcra, addnimechtar, da 
cenel; hence from m-stems — danorpe, da llae, indagne, ace. masc. 


88 EbeVs Celtic Studies. 

or neut. da sale (dat. sig. dit sailiu Incant. Sg. in Zeuss). A neuter 
da g (two g's) also appears, 710. The feminine a-stems agree on 
the other hand with the Skr. -e, Slav, -e, Lith. -z, for they show 
the after-action of -i, -% (still preserved in di = dve, Welsh dui) : 
nom. diflisc (sing, flesc. ■=*flisca), di hudir, ace. di rainn, di drim, 
di persin, indibrethirso ; from za-stems nom. digutai, diguttai, 
di mili, Z. 315, ace. indiguthaighthi airdixi, 966. The i- and u- 
stems appear to have simply lengthened the end- vowel. This was 
of course followed by a subsequent shortening, and then a drop- 
ping of the lengthened vowel : hence nom. masc. da preceptoir from 
-on, -Sri, da atarcud from -idu, -idic (gen. sing, attaircedo, nom. 
attdrcud like spirut, gen. spirito, spirto, spiruto), ace. masc. danog, 
ddnog from *ndgu, masc. or neut. indarecht from *rechtu, ace. 
neut. indd errend from -randi ? 

The in of the article consequently arises in the nom. and ace. 
masc. and neut. from *inna, *innd; in the fern, from *inni, 
*inni. It appears to have penetrated in the other cases in the 
same way that in the Greek -oiv has done in the genitive, or rw 
in feminines ; the frequent interchange, in the Irish, of the dat. 
and ace. after prepositions, is also to be taken into account, as 
well as the dying out of the cases which has been observed in 
Modern Irish (p. 80). The in cannot be well explained, organi- 
ically, at least, in the gen., in the dative not at all. 

Very few dual forms of consonantal stems have been unfortu- 
nately preserved. Of these the nom. da druith, and ace. da sligid, 
agree best with the Greek -e, for a Sanskrit -du, or a Vedic -a, 
would have led rather to druad and sliged. Nom. da thene, ace. 
da are, nom. ace. da ainm, da nainm appears to be decidedly 
inorganic. The frequent coincidence of the form with that 
of the nom. sing, has here, no doubt, brought about the invasion 
of the singular forms. 

In conclusion, it should not be forgotten, that in the Kymric 
not only are the commencing consonants in the substantives 
softened in the dual, but likewise in the following adjective, 
which is a proof that here also the nom. and ace. dual ended 
primitively in a vowel. 

§.11. On the Article in Modern Irish. 

In the modern Irish article an, about the relation of which 
to the old int, ind, I could not hitherto come to a satisfactory 
conclusion, I now recognize, with certainty, an intrusion of the 
neutral form, as the most colourless and weakest, precisely as the 
Middle High German had formed to its neuter daz a masculine 
and feminine der, diu, and the Lithuanian and Slavonian (to 
to) its tas, ta, tu, ta. The English use of that (pronoun) and 

On Declension in Irish. 89 

the (article) for all genders is especially important in this re- 
spect. 70 

It is a fact worthy of attention, but one hitherto scarcely 
noticed, that, besides the coarser, I may say the material, action 
of languages upon one another, which shows itself in the evi- 
dent borrowing of words and forms, a finer, a more spiritual 
influence is exerted. Thus, certain words, without being bor- 
rowed, are preserved living and active, by the neighbourhood of 
other languages, and some forms of thought and sound, words, 
expressions, conversational phrases, are so to say, indigenous in 
the soil. A comparative syntax would bring many examples of 
this kind to light, especially in the languages which have grown 
up on Celtic ground, and might determine how much may be 
ascribed to accident, and how much to intellectual influences. In 
the Phonology, for example, the Kymric id, oi, representing the 
Gaedhelic e (even in loan-words like cera, W. 2. kuyr, 3. kwyr, 
Cornish V. coir, Armoric coar) is parallel with the French oi, 
representing the Latin e (avoir = habere) ; again, the Celtic action 
of the final sound on the following word has a parallel in the 
transporting of the final s to the next word in les amis, etc. 
Among the words and word-forms which have been preserved on 
Celtic ground, we may mention: English, witness = Gaedhelic 
jiadnisse (testimonium), and the English names in -ton, along 
with the Gaulish in -dunum. Of importance in Syntax are : the 
French intercalation of the pronoun in je faime, je ne faime pas, 
as in both branches of the Celtic ; the French cest moi and the 
English it is me = Gaedhelic isme; the English leaving out of the 
relative in, the man (whom) 1 saw, as in the Gaedhelic. 71 Now, 
in this respect the English that, the, for all genders, are not 
without importance for the Celtic also, and permit us to conclude, 
that in the Modern Irish an fear for the Old Irish in fer, an 
analogous process has taken place. — The relative an (a, no, n) 
appears to belong to the same stem ; we may compare the fluctua- 
tion between the relative and the demonstrative in the Homeric 
language, the peculiar use of the Old Persian hya, which Bopp 
also, 72 as I myself did, 73 now looks upon as an article, and the 
German antiquated relative so. 

70 [xxxi. This is an ingenious error. The neut. article is quite lost in 
Middle Irish, and the Modern Irish article an (an t before a vocalic anlaut), 
bears the same relation to the Old Irish in (int) that the Modern Irish preposition 
an (written a n-) does to the Old Irish in ; or the Modern Irish interrogative par- 
ticle an does to the same particle in the Old Irish, viz., in. But here, as elsewhere, 
more is to be gained from Ebel's mistakes than from many another man's truths. 
The relative an, a, is doubtless identical in form with the neut. article =*sa-n. 
Ebel has since corrected this error. See, infra, On Phonology in Irish, §. 2. p. 138.] 

71 [But the two last named constructions are found also in the Scandinavian 
languages, where no Celtic influence is possible.] 

72 Vergl. Gram. I. 473. 2nd Ed. 73 Zeitschrift f. Vergl. Sp. v. 305. 

90 EbeTs Celtic Studies. 

§. 12. On the So-called Prosthetic n. 

[The term Prosthetic w, used by Zeuss, is what Irish grammarians erro- 
neously call an eclipsing n. Mr. Stokes in the papers above quoted, and Dr. 
Ebel here show that this n, in the majority of cases, belonged to the word imme- 
diately preceding that to which it seems prefixed.] 

Mr. Stokes, in his valuable observations on the Irish declen- 
sion, has agreed with my remark, that the n of the inflexion has 
been preserved in teora nguttae, and here and there also be- 
sides the n of the article, and has communicated several examples. 
Zeuss, curiously enough, has altogether misunderstood this n, 7i 
and everywhere looked upon it either as a superfluous addition 
or as a shorter form of the article, e. g., before aile, although 
there it appears only in the nom. neut. and ace. sing, and gen. 
plur. of all of the three genders, — often in combinations where 
no article is possible. As a relic of the article I Lave met with 
this w, only in very few places, and then as the remains of the 
shortest forms: an (a-n-) in tresngne, Z. 611, where the E of 
tres still indicates an a dropped out, and ni epur nIsin (non 
dico hoc, instead of anisin) 352 ; in (ace. dual) in etarndi- 
rainn 278, 614, probably as gen. dual in cechtarnai, nech- 
narnai 369 (compare the plur. innan Ai). The n in lasin 
ngutai (instead of lasinn gutai) 619, 1017. The most of the 
other examples are clear enough. I shall give here some proofs, 
which may easily be increased. Nom. and ace. neut. folad waill, 
OLCC WAILL, DES.(i.e., desimrecht) waill, pronomen waill 363, 

IMB&LRE ftAILL 580, MOR WAMRI 596, 889, GRAD 7ZEPSCUIP 1048, 

am. nach annse tiduib (ut non difficile vobis) 703, huare 


tiab (previously, ilsenman) 367, dered ^betho 985, is-fuath 


Ni ARMISOM archumactte ([nam] potestatem nequam non nume- 
rat ipse pro potestate) 247, nIfail nach waiccidit (non est 
ullum accidens) 1016, nicumscaichthi cumacht^: wairi (non 
mutanda potestas propterea) 1015, n! fitir imorro olc wetir 
(nescit autem malum omnino) 1003, laa mBRATHA 479, allaithe 


extremo (ace. temp.) in fine] 316, isnoichtech re wiuil (est 
undetricenale spatium Julii) 1075, isgnath gao et f!r wand 
359. So also — arindi atreba toxal wand 359? Ace. masc. co r!g 

74 [xxxii. Not so. See Zeuss G. C, page 263, where he conjectures that 
the very form cited here by Ebel, teora h, may stand for teoran.'] 

On Declension in Irish 91 

h ilainglech Colman's hymn — Lib. Hymn. 10 (to the many 
angel'd king), according to a friendly communication of Mr. 


fiADAROE 481, infogur nisiN 1014; without the article, besta- 
tidwisin 611, aes wesci 1074 (three times), nifail chumscugud 


leth TiGOTHO 1013 (consequently leth is also masc. like recht), 


616. So also no doubt: nach waile 368, toiniud ?ziressach 229, 
nert wainmnedo 975, attlugud mBuiDE 1048 (the ace instead 
of the dat. ?), cach^oen crann 999? I am not quite certain of 
the gender in, fri cumtach wecolso 260, cumtach tiirisse 


neuter) 56. Ace. fern, fricach waimsir 367, cech waidche 
(instead of aidchi) 888, isarnach windocbail moir 262, hi 
cach tzdeilb 7 hi cach tarmorcenn 367 (translated by Zeuss 
as the dat.), i persin tiaili 363, frirainn tiaili 608, cen 
gutai wetarru 1017; also doubtless, roscarsam frib denus 

mBEICC 310, HIRES 71ABARCHE 229, SERC 7ZDEE 55 (just as 

nem, delb occur in the ace), cen alpai 74ETARRU 616, 75 frial- 
pai rcDESiu 595. Gen. plur. masc. innamball waile 229 ; fern. 



Some spurious prepositions, it would appear, may be recognized 
as accusative forms by the n, most distinctly taresi in — u. tar- 
hesi ni (u for i) 1012, olcc taresi nuiLCC 617, but indegaid 
also in — indegaid wde 619, indegaid wgutt^ 1013, and do- 
chum in — dochum itBEE 620, dochum niRissE 461 (bis). 

The n of ainm-n belongs to the stem in — ainm wafstil 229, 


doberr ainm jidoib 45 7. 76 According to this my observation 
(p. 65), " probably derived, however, from m i and not a primi- 
tive w", must consequently be cancelled, and the single example 
with an aspiration ainm thrIuin Z. 249, considered as an irre- 
gularity. 77 As yet I have failed in finding for the masculine 
and feminine w-stems an example of the aspiration, or of a morti- 
fied s, f; I have also, however, nowhere found an n ; it conse- 

75 According to Stokes (Beitrage zur vergleichenden Sprachforschung I. £68) 
the n of alpai-n and inrindide-n belongs to the stem. 

76 See last note. 

77 [xxxiii. The n in ainm rcapstil does not belong to the stem, but (as in 
pronomen naill cited by Dr. Ebel himself, supra) is simply an example of the 
natural tendency to prefix after all neuters in the nom. and ace. sing, an n (m 
before b) to the following adjective, if this begin with a vowel or a medial.] 

92 EbeVs Celtic Studies. 

quently appears as if the neuter only preserved the N as in the 
Latin and Slavic, — *anmen like nomen and ime, while the mascu- 
line and feminine dropped it,— *brithema like homo and kamy. 
The wis much less clear in cechtarnai, nechtarnai Z. 369 
(which I consider to be a relic of the gen, dual of the article in, 
on account of dochechtar nhai, evidently the dative, and of the 
genitive plural innan ai), sliab nossa 888 (perhaps ace. ?), sirid 
inrindide nuile (see note 75) 366, 586, arbertar as noen 


naffracdai 676, where it appears to be in part actually er- 
roneous; cotIr nerend 74, appears to indicate a change of 
gender (comp. recht, leth, nert) ; even theie, however, Zeuss 
also gives fir nerend (viri Hibernise) with an enigmatical n. 
There is perhaps a threefold preposition do-air-in contained 


same way that con became mutilated in frecndirc ecndirc. 

But, very strange, the n appears very often after verbal forms ; 
mostly, perhaps exclusively, in dependent sentences, frequently 
after the so-called relative — aswoindae inspirut 360, aswed 

675, AM. ASftE ASSPLENDOR 333, AS72IRESS 456, AS720IPRED 476, 

do 229, aswdirruidig[the] anainmsin 265, ammi weulig 252, 
consechat wulcu 457, ata wanman sidi 894, ni cumcat ca- 
maiph ille 7 iste beta naithfoilsigthecha dondi as ipse 
667, intain bes winun accobor lenn 603. 78 

Notwithstanding that several examples still remain unex- 
plained, the vast majority show quite clearly, nevertheless, that 
the n is prosthetic, if at all, only in exceedingly few cases ; espe- 
cially the forms assumed by Zeuss, naill, naile, naili, nisin, 
nIsiu, and nand for and decidedly fall to the ground. 

13. §. On the Degrees of Comparison. 

Among the consonantal stems we have not mentioned the 
interesting ~n$ stems, the comparative, because no declensional 
forms of them are any longer to be recognized, with the excep- 
tion of adverbial dative forms, which offer nothing peculiar 
(immou, magis, indoa, minus indlaigiu, minus, intserbu amarius, 
indluindia commotius). As in the accusative plural, the 
primitive -ans has split itself into -a (consonantal stems, femi- 
nines, and the article) and -u (masc. a-stems), so here also we 
find both forms, the ~a in the more ancient, the -u in the newer 
secondary formations. Of the former rnda with its parallel forms, 

78 May it be, that as in Greek, an v tyekKvariKov existed ? Stokes also compares 
ammi-h with eafiev. 

On the Degrees of Comparison. 93 

corresponds to the Lat. major, Goth, mais, maiza; the Kymric 
form, W. mwy, Corn, moy, Armor, muy, which deviates somewhat 
in the vowel, has still preserved the j, i, and like all similar 
forms, has thrown off the final vowel, together with the s. Oa 
(minor) appears to have been formed after the superlative oam = 
Skr. avama, instead of Skr. avara, therefore properly : inferior, 
deterior; nessa = W. nes has been already several times com- 
pared with the Gothic neJiv nehvis, its superlative with the Osc. 
Umbr. nesimo, and the dropping of a guttural surmised ; tressa 
(fortior) — cf. W. traJia (audax, fortis) — exhibits the (in Sanskrit) 
regular throwing off of the suffix before the comparative ending, 
in opposition to tren, instead of tresn? (just as mdo along with 
mar); messa (pejor) appears to find its positive in the prefix mi- 
(Z. 833) = Goth, missa, although the latter aspirates the follow- 
ing consonants; in this respect, however, it has a companion 
in du-, which certainly represents the Skr. dus-, Gr. dvg-. The 
ss of the last examples appears to have arisen from sj, just 
as rr in ferr (melior) = Kymr. guell, gwell, whose Oscan and 
Teutonic affinities are compared in the Zeitschr. f. v. Sp. VI. 
421, does from rj, (compare also Skr. variyas, Gr. apdwv?). 
Lia (plus) has been elsewhere compared 7 ^ with the Greek 
ttXuujv, and ire shown to be a comparative. 80 The only compa- 
rative of that kind, which has joined itself to the second forma- 
tion in the Gaedhelic, laigiu or lugu (minor) — W. Uei, has re- 
mained true to the first — places itself alongside the Skr. laghiyas 
= Lat. levior but Gr. l\a<j<jwv; the substantive lagait (parvitas) 
is derived from this adjective. In the same manner the Gaedh. 
siniu = Lat. senior, and the Welsh hyn, deviate from one another. 
Among the Kymric forms liwy (longior), is (humilior), uch (al- 
tior), ieu (junior) = Skr. yaviyas which exhibit the rejection of 
the suffixes of hir = Gaedh. sir, isel, uchel= Gaedh. uasal, are 
particularly interesting. 

The second form u- is evidently only a contraction from -iu, 
in the same way as daltu occurs in the dative of the ia- stems 
instead of daltiu, maccidondu in the ace. plur. of the same stems 
instead of maccidondiu, and ditu, tichtu, epeltu in the nom. sing. 
of the £m-stems instead of ditiu, tichtiu, epeltiu; for the same 
reason laigiu and lugu, uilliu and oillu (plus), toisigiu and 
toisechu (prior) appear side by side. The majority of stems 
follow this formation, namely all derivative ones, hence isliu, 
huaisliu instead of the Welch is, uch. In the Kymric -ach corre- 
sponds to it (with retention of the s as ch), while the Gaedhelic 
-a has fallen off in the Kymric. The superlative in the Gaedhelic 

79 [See his paper, infra, " On the Loss ofp in Celtic", p. 161.] 

80 [See the paper referred to in note 79.] 

94 EbeVs Celtic Studies. 

separated into -am and -em, does not distinguish itself in the 
Kymric -am. How are these different forms to be explained ? 

In the first place let us recollect the double formation in the 
Teutonic and Slavonic, which have been already compared with 
one another in the Zeitschr. f. v. Sp. v. 309 sqq.; as Goth. -iza. 
Church- Slav, -ii, so does the Gaedh. -a belong almost exclusively 
to the defective comparatives, as the Goth, -oza, Ch.-Slav. -ei, so 
does the Gaedh. -iu attach itself to all secondary formations. We 
know further that j disappears in every position in the Gaedhelic 
and at least in the middle in the Kymric, except where it is 
preserved as the vowel i. Finally, we found (p. 54), that the 
ending ~e (dat. -iu) not only represented the Skr. -ya, but also 
frequently -aya, an origin which is still sometimes marked by 
the writing -ae (gen. -ai). Now as the Kymric forms (and also 
the single Gaedh. form ferr), as well as the analogy of the 
Teutonic and Slavonic, compel us to assume a shorter form 
than the Gaedh. -iu as the basis of the Gaedh. -a, the following 
hypothesis may best recommend itself: — The Celtic formed two 
kinds of comparatives, an older one in -jans, a newer one in -ajans 
(-aijans?-djans?); the -ns fell off in both in the Gaedhelic, as in 
the ace. plur., and left behind a long vowel, which was afterwards 
shortened ; the^" dropped out in the first form, but left behind some 
traces in Gaedh. messa, tressa, ferr (.?), in W. mwy, llei, hwy (?); 
in the second, -aja contracted itself to -a in the Kymric, and -aj 
to -e, i in the Gaedhelic ; in the Kymric, the s remained as ch 
after the (primitively long?) a, as in chwech six = Gaedhelic se, 
but fell off with the vowel in the shorter form. If this hypothesis 
be correct, the e (i) of the Gaedhelic superlative (stem -ima, nom. 
-em, gen. -im) must have been shortened from e or i. 

In Modern Irish, -iu has become e (passing through i?) as in 
the nominatives of the w-stems, e.g. laige from laigiu, like naoidhe, 
from noidiu (compare Pictet, Beitrage I. 83). 81 

81 [The following is the passage referred to. " Eriu, as it appears, is a still 
more ancient form of the nominative and accusative. It is so in two quotations 
by O'Connor from Eochaid's poems (belonging to the ninth century — Proleg. 
II., 40, 42). 

H Eriu oil ordnitt Gaedil, — Hiberniam totam ordinavit Gadelius ; H Eriu 
con huail con idnaib, — Hibernia cum gloria, cum armis (probably the accusative 
which cannot be determined, as the continuation is wanting). For this ending 
iu instead of the later e, compare in Zeuss (Gram. Celt., 268), noidiu, infans, 
Jrescsiu, spes, deicsiu, visio, ermitiu, reverentia, etc. (later and now naoidhe^ 
fresce, deicse, airmidhe), all of them having n, nn in the oblique case, where how- 
ever the u has disappeared, noiden, infantis, deicsen, visionis, etc., as in Erenn, 
Eireann. The proper name Fridriu or Frigriu (Ordnance Survey of Ireland), 
poem of Aileach, 1. 40, 43), whose genitive is Fridrenn (39), Frigrinn (1), Frigrind 
(2, 53) is another example".] 



§. 1. Views regarding the special affinities of the Celtic, and 
words borrowed from the Latin, 

THE European members of the Aryan family of languages 
form a chain, both ends of which reach over into Asia. The 
Greek undoubtedly shows the most numerous points of contact 
with the Asiatic tongues. On the other hand, however, the 
Slavonian exhibits the greatest number of special agreements 
with the Iranian. 82 So in like manner the neighbouring links, 
within the chain itself, are universally acknowledged to be the 
most closely connected with one another, Greek and Italian, 
Slavonian and Lithuanian, Lito-Slavonian and Teutonic. Very 
naturally, therefore, in the Celtic also, winch lies nearly in the 
middle between the others, most points of contact will be found 
with the Italic on the one side and the Teutonic on the other, 
and through both with the other already established branches of 
the European division. It is, of course, very difficult to decide 
to which of the two it stands nearest, perhaps even impossible 
at present, when, as regards the question of the separation of 
languages, so much is still debateable, and when a comparative 
syntax is still quite wanting. 

Lottner has pronounced in favour of the " Northern"; Schlei- 
cher for the Pelasgian languages; both, notwithstanding the 
divergence of their views in other respects, agree in placing the 
Latin nearer to the Celtic than to the Greek. I look upon so 
much only as proved, that the Celtic stands closer to the Latin 
than to the Greek. The circumstance that feminines in -oc, -us, 
only occur in the Classic languages appears to me to prove con- 
clusively a nearer connection of those languages with each other 
than with any other. In other respects also I have not adopted 
Lottner's view in so wholly unconditional a way as Schleicher 
appears to assume ; but, on the contrary, I have often expressly 

82 To this category belongs, besides many other points that Schleicher has 
brought forward, also the frequent occurrence of the suffix -ka where it is foreign 
to other tongues, e. g., Slav, sladuku (dulcis) = Lith. saldus, like Old Persian 
vazarka (magnus) = Skr. vrhat. 

98 EbeVs Celtic Studies. 

indicated, in general as well as in particular, the points of 
contact of the Celtic with the Classic languages. This shall 
not, however, deter me from exposing the points of agreement 
of the Celtic with the northern tongues, in accordance with his 
request. Only I shall take the liberty of beginning with a 
point which Schleicher designedly left aside, namely, with the 
vocabulary of the Celtic, partly in consequence of the accidental 
direction of my studies, partly in order to at once meet a fore- 
gone conclusion, that many (although certainly not Schleicher) 
might draw from this very source in favour of a closer relation- 
ship with the Latin. 

Indeed, it appears at first sight as if the Celtic languages had 
an especially large number of words in common with the Latin. 
If, however, one looks closer, by far the greatest number (even 
in Old Gaedhelic, but still more in the Kymric dialects) are 
seen to be foreign or loan-words, often so deceptively assimi- 
lated, that when one is about to take off the mask, he involun- 
tarily draws back his hand. So says, for example, Zeuss 
(p. 80) : Non tanguntur certe tenues in vocibus peregrinis re- 
ceptis. But words like accidit (t=n£), ethemlagas (th=i) which 
yet, unquestionably do not belong to the earliest loans, show 
what little claim this rule can have to general validity. At 
most, we can only judge thereby of the greater or lesser perfec- 
tion of the appropriation and mastery of the foreign material, and 
in this respect the Kymric surpasses almost the Gaedhelic. 

I have constructed two glossaries from the Grammatica Celtica, 
of which the Old Gaedhehc one may be considered to be pretty 
complete ; the Kymric one may, however, undergo considerable 
enlargement. From these I give the following lists, as the foun- 
dation of further glossarial researches. In order to supply in 
some degree at least the want of an organic orthography, I have 
retained here also the method of denoting the Kymric dialects 
which I adopted on a former occasion; 83 unmarked words are 
Old Gaedhelic. I did not like to pass over the Kymric, because 
the peculiar phonetic relations of the two branches of the Celtic 
family necessitate a mutual supplementing and explanation; 
besides, loan-words are often accidentally wanting in the old 
documents of the one or other dialect in Zeuss. In the case 
of the loan-words, a peculiar difficulty occurs in the Kymric 
dialects, as it is often scarcely possible to decide whether a word 
has been introduced directly from the Latin, or through the 
French, Anglo-Norman, or even the English. The decision 

83 Beitr., I. 427. [See Appendix II., p. 184, for an explanation of the abbre- 
viations used in the following lists.] 

On the Position of the Celtic. 


is the more difficult because the Kymric vowel-changes mostly 
agree with the French (especially in the treatment of S) ; and 
because in the case of the dialects with regard to which the 
idea of such a medium would first arise, Cornish and Armoric, 
our sources are too modern to help us to answer the question from 
the chronological side. In order to avoid, as much as possible, 
mistakes in this respect, I have only indicated the medium 
where it appeared to me certain. The Old Gaedhelic, especially, 
contains a great number of Latin words (or of Greek ones bor- 
rowed through the Latin) from the domain of the Church and of 
science ; but others also are not wanting. 


abbas=[a&], V. abat, W. 2. pi. ab- 
ba deu. 

[abecedarium, aibgiter, W. egicyddor.~\ 


&ccent\is=aiccent aiccend, d. aicciund; 
[W. 3 acen.~] 

accidens=accwfrf aiccidit n. 

facer, aicher, W. 3 egr.~] 

acetum : dctegim (aceo). 

[actualis, achtailJ] 

acutus=t/c«^; noacuitigjide (acuenda 

adj ectivum— at/iec?, adiecht. 

adorare : adras (qui adorat), adrorsat 
(adoraverunt), adrad (adoratio). 


[adversarius Mid. Ir. aibhersedir.~] 

altare=a/to'tV f., V. altor ; [W. 3. 

altum=W. 2 alt allt all, 3. allt (collis, 
acclivitas, scopulus), V. ah (litus). 

anachoreta=V. ancar. 

ancora=i/?#or, V. ancar, [W. 3. angor], 
(the i as in ind- (dvn) and imb- (dfityi) 
and g like d in ind- (Gaul, ande-) make 
the borrowing a little doubtful ; the 
borrowing of the same word in other 
European languages (compare for in- 
stance Lith. inkaras), and of other 
naval expressions in the Celtic, speak 
however for it). 

aDgelus=aw^eZ aingel, "V. ail, P. eyll el, 
Arm. ael el hel. 

aninial=W. 3. aniueil, pi. "W. 2. any- 
ueilyeit, 3. anniueileit aniueileit anni- 
ueilet, Arm. aneualet. 

apostohis=apstal, V. Arm. apostol, P. 
pi. abestely. 

applicare : Arm. em em applicquet. 
(applicate vos). 

[aratrum, arathar W. 3. aradr.~\ 

argentum=[azr<7e£, argat], W. 2. ariant, 

3. aryant, V. argans, P. arghans, 
Arm. argant. (External evidence of 
the borrowing is no doubt wanting, 
the evident borrowing of the name 
of gold, as well as pbysico-geographi- 
cal reasons, speak however in its fa- 
vour). [? cf. Gaulish Argentora- 

(?) anna — arm, arma, d. isind-airmm 
(in armatura), W. 2. 3. arueu arfeu, 
P. arvow. 

arniilla=W. 1. armel. 

[ars, art.~\ 

articulus=a rticol, geu. sing.=n. pi. 
articuil, d. artucol. 

asinus= [asan~], W. 3 assen, V. P. asen 
(The grounds for the borrowing are 
elsewhere given). 

atomum=afo?72, in the phrase 7 unga 
7 atom (et uncia et atomum), not 
recognized by Zeuss, 312, 1076. 


[Augusti (mensis), augaist, W. awst.~\ 

aurum=or (gen. air), W. 2. 3. V. eur, 
P. owr, Arm. aour. (The r undoubt- 
edly indicates borrowing, cf. Sabin. 
ausum and Lith. duksas. Grimm con- 
sequently errs in his Geschichte d. d. 
Sprache 1027;. 

[baculum, bachall, W. 3. bagl.'] 

[balbus, balb.'] 

[baptista, bauptaistJ] 

bwptizo=baitsim ; ace. baithis, dat. 
baithius, W. 3. bedyd m., Arm. badez 

barba=W. 3. baraf, baryf; V. barf 
baref. (The borrowing is no doubt 
remarkable, I cannot however ex- 
plain otherwise /in contradistinction 
to Lith. barzda, Slav, brada, O.H.G. 

[barca, bare, W. barg.~] 


Ebd's Celtic Studies. 

M. Lat. baro=TV. 8. barwn. 

[basilica, baislic] 

[basium, bay, P.] 

battuere : V. bat (nuniisina\ TV. 2. Y. 

bathor (numularius, trapezita), P. 

bat ales (proeliari). 
[beatus=Mid. Ir. biat] 
benedico : headaches (benedicis), /lo&Sen- 

dachat (salutant vos), indatbenda- 

chub (benedicam te) : bendacht, TV. 

3. bendith, Arm. bennoez (benediction ; 

W. 2. bendicetic, P. benegis, Ann. 

ben(n)iguet (benedictus). 
bestia=TV. 3. bwyst-uil (appears to be 

compounded with nuZ like the Ger- 
man Maulthier, etc.), 0. Gaedh. 

be'isti f. pi 
blasphemare (Fr. bIamer)=T?. blamye. 
[Med. Lat. brace, braich, W. 3. and Y. 

[braehiuni, brace, Y. brech. W. 3. 

brassica=5 /-aisecA. 
brevis (svllaba)=6/-e/6. 
broccus brocchus (see Diez). Fr. broche 

=Y. broche (spinther). [Mid. Ir. 

[bulla, boll, TV. bwL] 

buxus=Y. box. 

ealanms=[TV. 1. colaiRemiou],W. 3. fee- 

Jeu-yn m. (singulative); V. kalagueli 
(stra men turn). 

M. Lat. caldaria=W. 1. callaur, Y. 
caltor. [Arm. &ao?er.] 

[ealhdus. callaid.] 

calix=[cai/ec^], Y. kelegel. 

[camisia, ca//«5€.] 

[cancella. eai'/^e/.] 

cancellarius=\V. 2. kaghellaur, kyghel- 

cancer=TV. 1.. Y. cancher. 

candela=Y. cantuil, W. 3. canmcyll : 
candelarius=ca/H^/o';V, Z. 744 ; can- 
delabrum : Y. cantul-bren. 

canon=acc. candin. 

[capellanus. Mid. Ir. cabellanacht.'] 

capistrum=TV. 1. cepister, 2. kebyster, 
pi. kebi/sireu kelesteryeu. 

[capitulus, caiptpl.~] 

[captus. cacht, TV. 3. cae^.] 

[caput, cant.'] 

carbimculus=ca/-7«oco/ Z. 1163. [TV- 3. 

carcer=e«/rar (gen. pi. earere, dat 
carcdir); TV. 3. karchar. 

(?) caritas (cf. charite)=K. * cardaut 
(beneficmnO in TV. 3. cardotta (jnen- 
dicare\ Tne ending -taut (==tdt) 
occurs especially in loan-words. 

car(o)enum instead of car(o)enaria 

=Y. ceroin (cupa), [TV. 1. ceroenhou, 

gl. dolea.] TV. 3. herwyn (lacus. le- 

[carpentum. carpat.~] 
pasesas={cdise\, TV. 2. terns, 3. caws, 

Y. caws cos. 
castellum=: [c a is e/]. pi. W.2. c est ill, 2. 

[castus. cast; castitas, castoit.] 
catena=W. 3. cadwyn. 
cathedra=TV. 3. cadeir (sella), Arm. 

[cathohcus, cathlac] 
[Med. Lat. cattns=V. hat, TV. 3. cath.~] 
[caucus, cuacn, TV. 3. cawg.] 
caoles=V. caul (olera). 
causa=dat. cdis. 
[cedria. cedirJ] 

[cella. cell, Y. tal-gel gl. cellarium.j 
census=cw (census, fiscus. vectigal . 
cera=[ceiV], TV. 2. kuyr, 3. kwyr,Y. coir, 

Arm. coar. 
[cervical. ce'rchaill.] 
[cervisia, ceirbsire, ••brewer".] 
character=camcfaar n. (littera). 
chorda=P. pi. kerdyn (funes), V. cor- 

den (fidis). 
christianus=[cresen], TV. 3. cristatpn, 

Arm. christen. 
[chrisma, crismaL~] 
[ciheium, cilic] 
circare (see Diez)=TV. 3. kyrchu, Arm. 

querchat. querchit (qmerere, per- 

gere, intrare). 
[eircinus, cercenn, TV. 1. circhinn.] 
circulus=acc. cercol. 
civitas=W. 3. kiwtawt; kiwtawticyr 

clarus=P. clear, Arm. scler (with pre- 
fixed s). 
[classis. dais.'] 

[claustrum, V. clauster, cloister.'] 
clericus = [clerech], Y. cloireg, Arm. 

[clima. climata. pi.] 
M. Lat. clocca=[c /occ,],Y. cloch; clechir 

(tintmnabiihmi), clechti (cloccarium), 

clochmuer (carnpana\ 
coccus: TV. 3. coch (ruber \ pi. coch- 

colonic TV. 3. kalaned pi. habita- 

(?) comrnba=co /«//?. V. colom, Arm. 

coulm, (cf. Slay. golq.oi in Schleicher, 

columna=TV. 2. kolouen (i.e. columina. 

with an intercalated vowel;, 0. Gaedh. 

[colomna. n. pi.], columnat (columel- 

On the Position of the Celtic. 101 

*cominitiare, Fr. commencer; Arm. [culcita, coleaid.~] 

comancc (initium), m=mm. cultellus=W. 1. cultel (" artuum"), 

commatres=Arm. coma(e)zreset. celeell ("culter"), 3. cyllell, pi. cyl- 

commodum—comad-as ; comadasogod leill kylleil (sica) ; V. collell (" cul- 

(accommodatio). tellus"), kellillic (" artavus"). 

[communio, Mid. Ir. command, W. 3. culter=W. 1. cultir, [W. 3. cwlltor], V. 

cymun.] colter. 

[Med. Lat., companium, companacht.~\ cuprum=V. cober. 

[compar, W. 3. cymhar.] [cymbalum, cimbal.] 

comparativus=co mparit, pi. -iti, gen. [cypressus, cuprisi] 

-ite. daemon=gen. demuin, gen. pi. demne. 

compatres=Arm. compizrien. damnare=P. dampnye, Arm. daffny ; 

concedere : Arm. concedis (consensi). P. dampnys, Arm. dqfnet, daffnet 

confessio=[coi6se], Arm. coffes. (damnatus). 

confligere : conflecktaigthi (congredi- [debilis, diblide, gl. senium.] 

endum). decedere=Arm. decedy. 

confortare=P. comfortye ; dyscomfortys decima=W. 2. decum, degum. 

(debilitatus, turbatus). defendere=Arm. difen ; [W. 3. diffenii], 

[consecratio, coisecrad; consecravit, V. diffennor (" excusator"). 

cusecar.~] denarius=a?j'rcaiY, Arm. diner. 

Consilium=[ci<m/], Arm., V. cusul, P. [deprecatio, diprecoit.] 

cussyl cusyll cusill. descendere=W. 3. disgynnu, P. dys- 

consona = conson, gen. consine. [W. kynna, Arm. disquennet. 

cyson.] desiderabat=P. deserya. 

[conucula (Med. Lat.), cuigel, V. kigel, despectus=P. dyspyth, Arm. despez. 

W. cogail.~] (li&bol\is=diabul, Arm. diaoul, P. pi. 

conventus=W. 3. hoveint m. (monas- dywolow dewolow. 

terium, fr. convent). diaconus=V. diagon; O. Gaedh. pi. 

[coquina, cuicenn.~\ bandechuin (diaconissae). 

coquus=[coic], W. 1. coc (pistor), V. dictSLtoT=dictatdir. 

kog (coquus); coquina=:W. 3. kegin, [dies Jovis, deyow, P.] 

V. keghin (the k proves the borrowing, [dies solis, dew suit, P.] 

the true Kymric forms have p : V. digamma=cfa7ai7ft. 

popei (pistrinum), peber (pistor), W. dignus=Arm. din (French or di- 

2. popuryes, pophuryes, f.). rectly ?). 
eorona=[Mid. Ir., corohi\^V. curun ; W. [diluvium, diliuJ] 

3. coronawc (coronatus). discere=P. dysky, Arm disquif; W. 2. 
corrigia=W. 1. corruui, 3. carreL desko (didicerit) ; P. dyskas (doc- 
(?)corylus=co/Z; W. 2. coll (coryle- trina). 

turn); col ennf^=V .col-viden(corylus). discipulus==c?esaJDwZ, V. discebel, Arm. 

craticula=W. 1. gratell. desquebl, pi. P. dyscyplys dyscyblon, 

creator = V. creador, Arm. croeer Arm. disguiblion. 

crouer ; creatura=V. croadur(?), [discretus, discreit.] 

Arm. croeadur. — W. 1. creaticaul discus = \jesc~], W. 3. dyscyl disgyl 

(genialis). (discus, lanx). [W. 1. disci.'] 

[credulus, credo,!.] divinator=Arm. diuiner. 

[crepusculum, crepscuil.] doctus=W. 3. doeth (prudens), pi. 2. 

[creta, criad.] doythion (sapientes), 3. doethon 

[cribrum, ribar.] (docti). Also Arm. doetaf (fallo) ? 

crudelis=Arm. cruel (French, or di- dolor=W. 3. dolur. 

rectly ?). [dominica (dies), domnach.] 

crux=crocA, V. crois, P. croivs. draco=[e/rac], W. 3. dreic, pi. dreigeu. 

crystallus [Mid. Ir., crisdal], trans- dubitare=Arm. douetaf; douet (du- 

formed in W. 3. krissant m. bius) ; doetanc (dubitantia). 

(?)cucullus (first in Martial and Juve- (?) durus=[dur], W. 1. dur (dirus), 3. 

nal)=[cocAw//] , V. cugol. — According dyrys (durus). 

to Diefenbach (n. Jahrb. f. Phil. u. ecclesia=dat. abl. aeclis, gen. ecolso 

Pad. lxxvii. p. 756) ; the Latin word ecilse, etc. ; W. 2. eccluis, 3. eglwys, 

had been already borrowed from the V.P. eglos, Arm. ylis. 

Celtic. eleemosyna=a/Waw., ace. almsin (erro- 


EbeVs Celtic Studies. 

neously given as a nominative at 

p. 59). 
elephantus=W. 3. elijfeint, V. oliphans. 
emendare : W. 2. emendassant (emen- 

episcopus=ejoscojo, V. escop, pi. W. 2. 

epscip, 3. escyb; archiepiscopus=V. 

archescop, Arm. archescob, pi. W. 3. 

eremita=V. ermit. 
esculus : escal-chaill (esculetum). 
(?) esox—[iach], W. 3. ehawc, V. ehoc ; 

the Latin was perhaps borrowed from 

the Celtic). 
French, estonner, etonner=Arm. esto- 

excommunicatus=W. 2. yskumunetic, 

according to Z. also=eskemun ; 3. 

ysgymunn (maledictus). 
evangeliurn=V. geaweil, Arm. auiel? 
faba=sei&, (cf. frenum, flagellum). 
facies, Engl. fa,ce=B.fetkfyth. 
fagus=W. 3.Jfa,Jfa-wyd. [0. Ir.fagde, 

faginus, Z. 765.] 
[falco==-V.falhun, Arm. faWhoun.'] 
fallere : P. fall (defectus), fyll (deest), 

fallens (deficiunt, peccant), Arm. fall 

(malus), V. guin fellet (acetum, i. e. 

vinum corruptum). 
favere (faustus): V. Jodie (felix). 

fenestra^V./enester. [Mid. Ir. sinistir.'] 
ficus=V. Jic-bren ; O. Gaedh. Jiculdae 

fides=P. fethfyth, Arm. fezfeiz. 
£.g\irsu=ind-Jigor (figuratio). 
finis=W. 2. fin, P. fin-weth, Arm. fin- 

uez; finire=Arm. jf m'ssa/ (Romance, 

firmamentum=[M. I. Jirmamint.], V. 

flagellum=sro^e/?, W. 3. ffrowyll. 
flamma=W. S.fflam f., Y.flam. 
[flecto, slechtaim.~] 
foeniculum = V. fenochel. [W. 3. 

M. Lat. follis (cf. Diez-Worterbuch, 

where, however, the perfectly analo- 
gous German Windbeutel is forgot- 

ten)=W. 3.ffol, Y.fol, P. folfoll, 

Arm. foil (stultus). 
M. Lat. fontana=W. 2. finnaun f., 3. 

ffynnawn ; V.Junten; F.fyntenfynteon; 

Avm.feunteunfeunten, pi. feuntenyou. 
M. Lat. forestis foresta=W. 3. fforest 

m., Arm. forest. 
forma=V. fitrf. [W. 3.ffurf] 

fossa=W. 2. Arm. Jos, f oss. 

fragrare (with dissimilation) : V. flair 

(odor), Arm. flerius (foetidus)." 
frenum=snan, W. 1. Jruinn, 2. fruyn, 

3. ffrwyn. 
fructus=Arm. Jruez; W. 3. diffrwytk 

(sine fructu). 
fugere=W. 3.ffo; P. fo (fuga).— V. 

fadic (profugus). 
fulgur, French foudre=Arm./owfrr. 
funis=W. 1. pi. Juniou (vittae), 2. pL 

Junenneu (ligamenta). 
fur=V. fur (sollers, prudens), Arm. fur 

furca=W. 3.fforch. 
furnus=V. Jorn (clibanus). [0. Ir. 

fustis=W. Z.ffust (flagellum) ; ffustawd 

(pulsavit). [suist.~\ 
gentes=genti geinti (m. as in French) ; 

gentlide (gentilis); gen. f. geintlecte 

(gentilitatis) 1059. 
genitivus=g enitiu f. 
[gens, pi. geinti; gentilis, geintlide.~] 
[glossa, gluais.~\ 
[gradale, Mid. Ir., greddilj] 
gi8idus=grdd n. (gen. grdid), V. grat. 
ypd(p(i) : W. 2. gref (liber, chirogra- 

phum), W. 1 . grefiat (notarius). 
gratia=P. gras. 
[gratias agimus, grazacham.'] 
gravari=P. grevye. 
gravis (accentns)=graif. 
M. Lat. gridare (quiritare)=W. 3. gryd 

(clamor), grydiaw (vociferari), gridu- 

an (vociferatio). 
[habilis, W. 3. abl.~] 
haeresis = ace. innerese ; haeretici= 

heritic pi. 
[Med. Lat. hanapus=V. hanaf] 
(Fr. haster, hater=Arm. hastomp, fes- 

[hastula, asdul.'] 
[historia, Mid. Ir. sdairJ] 
[honor, onoirJ] 

(? hora=war, P. or, W. awr ?) 
hospes=W. 3. pi. ysp. 
humilis=[ww2af], V. huuel; humilitas= 

(Ji)u?nalddit (h)omalddit, V. huueldot. 
[hymnus, ymun.] 
idolum— icfo/ m. 
[idus, id.'] 
[imago, V. auain.~] 
(impedicare ?) Fr. empecher = Arm. 

ampeig (impedimentum). 
imperator=W. 3. amherawdyr, f. am- 

herodres ; V. e?nperur, f. emperiz; W. 

3. amherodraeth f. (imperium). 

On the Position of the Celtic. 


inccnsum (cf. Fr. encens)= V. encois 

(thus), incoislester (tlmribuluin). 
int'amis=Arni. iffam. 
infernum=t/Mrr»n, gen. ifirnn ; W. 2. 

ufern, 3. ujffern, B.yffarn yfforn. 

[initium, init, W.ynyd, '"Shrovetide".] 
[instrumentum, Mid. Ir. instrumint.~\ 
[interjeetio, iuteriecht.] 
i nterrare=Arm. enter raj. 
[V jejuiiiuni, aine.] 
j iidex : iug-suide, (tribunal). 
[jusculum=V. iskel.] 
justitia=Arm. iusticc. 
[kalendar, calann.~\ 
laicus [7(/(v//],V./< w, pi. W. 2. lleycyon. 

[W. 1. /<!cvc.s, gl. niaritae.] 
[latex=V. fad. ] 
latro •-■— V. lader; P. fader ladar, pi. 

ladron laddron; \Y. 3. /A A//- llevdyr. 
[lector, tei/toir.] 
[lectus, t< e/it.] 
legalitas=Arm. haltet. 
l&gGre=legend ; airlech (recita), inroleg 

(mini legit?) etc. 
legio : W. 3. kaer-Uion (eastra legio- 

leo=W. 3. Hew, V. leu. 
liber=&5wr Ichor, V. fruer; pi. P. 

luffrow, Arm. A iffriou. 
[ligo, W\ I., pi. tiuatt.] 
lilium=V. A'AV. 

(?) linum ftn (rete), K. ftn (linum). 
littera (not litera)- --titer ; W. 3. ////- 

thyr-en, V. lither-en (singulative). 
loculus==V. logel. 
locus=fac, Arm. lech, W. 3. //<- (in no 

case primitively related, as the 0. 

Lat. stlocus shows). 
[loeusta=V. legest, W. 3. llegest ' lob- 
longa (syllaba)=A>A///. 
(navis) longa=\V. 3 /Aw/ f. (navis), 

pi. 2. loggeu loggou, 3. Uongeu; W. 3. 

llyghes Uynghes (elassis); O. Gaedh. 

[/<>////], /(>/7<>m//\(navigatione),Z. 1 129. 
[lorica, luirech. ] 
lucerna = luacharnn, V. lugarn. [W. 

\umiris=l it na'ir. 
magister=nom. pi. magistir, ace. pi. 

magistrw, V. maister. l\ Arm. mester. 
(?)ma,ior~—[M(ier,i)H>r->iiuer],W. 1. 2. V. 

watV, W. 3. waer. 
maledicis = matdachae <• maledictio = 

maldacht; maledic == Arm. millic; 

maledictus=W. 2. mettdicetic. 
malitia : Arm. dima/iec, diuaticc. 
[malva=V. ma/o«.] 

| mancus==V. wm/^ (leg. mane ?), Arm. 

manere : Arm. manen (manebam). 
| manna, mainnJ] 

[<a?, V. mantel.'] 
[maims, man.'] 
margaritae=W\ 3. mererit. 
niartulus (inartellus)— W. 1. morthol 

(seta), 3. myrthiv (malleus). 
maxtjriwai===martre f. pi. martri; Arm. 

medicus=V. medhec, W. 3. medic ,• 

\Y. 2. medhecynyaet f.=V. medhec~ 

uaid (medicina); W. 3. medeijiny- 

aeilni (mederi). 
membra=mewi6ur, pi. 
memoria tin buir. 
mendicus : mindechu mindchichthiu 

(tenuior, properly mendicior), mind- 

eliitjitir (emendicant). 
mensa==[mios], V. mnis, W. 3. mwys (?) 

cf. Goth, mis, O.II.G. mias. 
mensura — Arm. musur t (c£.W. 1. doguo- 

misnram ? Z. 107G). 

meretriz —mertrech, meirddrech. 
metrum : gen. metair metir. — W. 3. 
metrut (cogitabas). 

miles = mit, W. 3. mitier; militia 

(?)miHe=wn7e f., K. /////(from mitia?). 
[millefolium. V. min/ct.] 
ministrare: Arm. ministren (ministrem), 

V. me tast ear (pincerna). 
(minus facereDiez) Fr. mesfaire, me- 

faire: Ann. mesjectouryen (malefac- 

M. Lat. mirare: Arm. mir (serva), miro 

(videbit), »«'ret=P. meras (servare. 

minis: P. mdrih (miraculum); Arm. 

maruaill (mirabile) -Fr. merveille. 
modus=moa' (gen. muid, dat. mud). 
(?)molina=-mu2ien7i (pistrinum), K. wc- 

ftn, pi. W. 3. melineu. 
monaehus=[01d Ir. and] V. manae/i, pi. 

W. 2. meneieh; f. W. 3. manaclus, 

V. mauaes. 
monasteriuni- gen. pi. mom's/ re. 
moralis: dat. moral '-its (praecepto). 
[mortieinium, muirtchenn.'] 
morus— V. moyr-bren. 
[Med. Lat. muito=moft V. mots.] 
(?)mulus=acc. pi. mulu. 
murus=[mMr], W. 3. mur, pi. muroed. 
mutil=miit ; mutU8=:W. 3. /nut. 
niyrias=W. 3. mijrtl. 
myrtus : mirt-chaill (myrtetum). 
natalicia=W, 2. nodolye, 3. nadolic 

(nativitas). [0. Ir., nottaie.] 


EbeTs Celtic Studies. 

[nates, ndt.] 

natio=Arm. nation. 

negotium=W. 3. neges f. 

neutrum=newfor, neutrdlde, 

[nimbus, nimb.~] 

notsu=not pi. 1011, nota, 1016; nota- 

Tius=notai?'e notire, Arm. noter. 
[novellus, W. 1. nouel.~] 
numerus=:W. 1. nimer, W. 3. Arm. 

niuer nifer, P. nevor. 
(? nuptise : W. 3. neithawr ?). 
obediens=Arm. obediant. 
[oblatio, oblann.] 
offerre : W. 2. of rum, Arm. oferen, pi. 

offer ennou (oblatio). \oiffrenn.] 
olea: ola-chrann (oliva), ola-chaill (oli- 

vetum) ; V. ohu-bren (oliva), — oleum 

=W. 1. V. oleu. 
[operarius, V. oberor.~] 
optativus=o/tfcuY optit. 
opus [opera?]=o6ar ? (saibes inobar, gl. 

inanem fallaciam Z. 1040) usually 

oipred, gen. oipretho ; P. o6er ; Arm. 

o&er auber, pi. oberou enffrou ; P. 

oberor (operarius) ; V. drochoberor 

oraculum: [aiVecaZ], oirclech (flamen= 

[orate, oraiV, W. 1. arawf.] 
[oratio, ace. s. orthain.] 
ordo=ore? orofc? or£ wr£, Arm. ?/rz; or- 

dmo=oi7-dni?nm ; Arm. ordren (ordi- 

natio), ordrenhat (ordinare). 
[ostiarius, oistreoir.] 
ostreum=V. estren. [W. 3. histr, Arm. 

[paganus, pagan.'] 
pagus=P. pow, U. Armoric (of the year 

833) pou. [W. paw.] 
[pallium, caille, W. 3. joalL] 
[palma=V. W. 3. pal/.'] 
palus=W. 3. pawl, pi. polyon. 
(palus) M. Lat. padulis (?)=W. 2. pull 

(fossa, lacana), V. pol (puteus) ; W. 3. 

pyllaivc (lacunosus, paluster). 
[panis, pain.'] 

papa=/)a/Jo, W. 2. pap, pi. papeu. 
papilio=/jMj0a#, W. 3. pebyll (tentorium, 

Fr. pavilion), 
[papyrus, joai/jer.] 
paradisus == [partus], Arm. paradis, 

parare=W. 3. peri (facere, jubere)? 
paries=V. pomnt {ui^=e instead of e as 

in the French paroi). 
[parochia, pairche, Mid. Ir. fairche.] 
pars=W. 1. part parth pard f., Y.parth, 

Armoric parz perz. [Irish pairt.] 
pascha=W. 2. 3. pasc, O. Gaedh. ace. 

caisc. f. 

[passio, pais.] 

patella=[W. 1. patel,] W. 2. padell f., 

V.padel-koern (" sartago") i. e. patella 

pauper=Arm. pi. peoryen. 
pausa (?)=r.W. 1. Arm. poues (quies), P. 

powesough (quescite) ; but W. 2. po- 

guis-ma, etc. (a place of rest), 
pavo— -W. 3. pawin, V. paun. 
pax=Arm. peuch. 
peccatum=£>eccac? m., W. 3. pechaut, 

Arm. pechet, pi. pechedou. 
pedester= W". 3. pedestyr (pedes). 
[rr'eXeKvg, W. 1. pelechi gl. clavae.] 
[pelliceus, pellec] 
pensus (Romance peso)=[p2ss], W. 3, 

pwys, P. poys (gravis, ponderosus). 
[pentecoste, cingcidisJ] 
peregrinus = V. pirgirin. [W- 3. pe~ 

perfectus, Fr. parfait : Arm. parfetaff 

[pergaminum, V. parchemin.] 
persona=persan, W. 3. person. 
petere=P. pesy, Arm. pidif pidiff"; 

Arm. peden, ph pedennou (oratio 

precatio) ; P. pi. pesadow — appetere, 

=Arm. appetaff. 
phial*=W. Z.ffiol, Y.fiol 
philosophus=/e/sM& ; philosophia=/e#- 

pethedic (minutus) W. 3. appears to be 

from the same stem as French petit ; 

its th points back to tt or ct. [Ir. pit, 

W. peth.] 
[pinnaculum, penakyll, P.] 
pinus=V. pin-bren. 
piper : [scipar.], W. 3. pebreid, pybreid 

pirus=-V. per-bren. 
[piscis=V. pise, W. 3. pysg.] 
[piscator=V. piscadur,~W .3. pysgadwr.] 
[plaga, plag.] 
plangere (properly planctare) = P. 

plentye (accusare). 
plenus : Arm. plen (omnino). 
p]ebs=0. Arm. (year 862) phi plue, 

plueu ; Sp. ploe ploue, pi. ploueou ; V. 

plui (vicus, parochia); Arm. ploeys 

(?) plicare=W. S.plycca; Arm. pliga- 

dur (voluntas, beneplacitum). 
pluma=[cZw?;z], V. pluuen (penna) ; "W. 

1. plumauc, V. plufoc (pulvinar) 
poena=pe1rc pian; Arm. poan (angustia), 

pi. poanyou; P. peynys (dolores). — 

Arm. penedour (amictione gravatus), 

W . 3.periydyaw(-poenitQie), O. Gaedh. 

pennit (poenitentia). 

On the Position of the Celtic. 


pommaille (Fr.)=Arrn. pomell. 
pondo=W. 1. punt m. 
pons— W. 2. pont, V. pons. 
(?) porcellus=W. 3. parchell, V. por- 

chel. [Ir. orc=porcus.~] 
porta, portus=/je>/^ m. (domus), Beitr. 

I. 334; W. P. porth m. pi. VV. 3. 

pyrth, P. porthow (porta). 
portare=W. 3. porthi (perferre), por- 

thes ; P. porthas (nutrivit) ; Arm. 2. 

porz (quaere, adjuva), porzit (subve- 

nite, sublevate) ; W. S.porthant (pro- 

visio, nutritio), porthmon (hospes, 

postilena=.\V. 1. postoloin. 
postis=:W. 2. post (columna). 
praebendarius=V. prounder. [pi. pron- 

teryon P.] 
praeceptum=p>-ece/?Z f. ; praeceptor= 

praedico = predchim, predach, predag ; 

Arm. prezec (praedicare). 
praelatus=[pre?atf], Arm. prelat. 
praeservare : Arin.j»?-eseryo(pragservet). 
praestare : Arm. prestis (praestitit). 
prandium— />/-o?W (prandere). 
[presbyter, cruimther ?] 
pretiare : P. praysys (celebratus). 
primus=/>>7/«, W. 3. prij-. 
princeps=P. prins, pryns, pi. princis. 
[prior, Mid. Ir., ba.n-prioir.'] 
prison (French) : Arm. diprisonet (ex- 
car ceratus). 
probus: amprom (improbus), amprome 

(improbitas), rondpromsom (q. id pro- 

bavit ipse), promjidir (probabitur) ; 

Arm. proffe, prouffe (probaret) ; P. 

previs, prefis (probatus). 
[prologus, prolach.~] 
pronomen=/)rono»ien n. 
[propositus, propost.~] 
propheta=V. prqfuit, pi. P. projusy. 
prudens=W. 3. prud. 
psalmus=sa///?, pi. sailm, ace. salmu ; 

psalterium=dat.safo'r, Arm. psaulter. 
[psalterium, saltair, gen. saltrach.~] 
[purgatorium, purgatoir.'] 
purpura=corcwr, W. 2. porffor. 
purus=jj?wr], W. 3. pur, purdu,purgoch, 

putana (Rora.)=W. 3. putein. 
(?) puteus=cwte, Beitr.1. 334 (strikingly 

reminds us of the Low German haute, 

kute, a pit), 
[quadragesima, corgais, W. grawys.'] 
[quaestio, ceist.~] 
[quinquagesima, cingices.'] 
[rastrum, rastal, W. 1. rasd.~] 

recommendare (Fr.)=Arm. recommant. 

? regnare=Arm. renaff— but compare 
Arm. roen (rex) — ? 

regula=-r2'«^»/, riagol ; Arm. reol. 

[reliquiae, reiUc.~] 

remus=/-a/ft (cf. Fr. rame), V. ndf. 

rendere for reddere (Rom.) : Arm. rento 

rete=V. i'uid, Arm. roed. [W. 3. rhiuyd.] 

rosa : ros-chaUl, ros-tdn (rosetum), 
rostae (rosarium). 

[ruta, V. rutc.~] 

[sabbatum, saboit, pi. sapati.~\ 


[sacrificium, sacorbaic.~\ 

sacrilegium=Arm. sacrihig. 

saccus=[sacc], V. sack. 

[saeculum, saigulJ] 

[saliva, W. haliw, O. Ir. saile.~] 

[salicastrum, saUestar, W. elestrJ] 

saltus=safr, gen. salto (astronom.). 

salutar&=Arm. saludomp (salutemus). 

salvare. Fr. sauver (with the old diph- 
thongal Norman pronunciation, see 
Diez. Rom. Gramm. I 2 , 425)=^P. saw 
(salva), sawye (salvabat), sawye (sal- 

sanctus=:[.s i a»cA<], W. Arm. sant, pi. W. 
2. 3. seint, Arm. sent. 

[Med. Lat., sappetus, V. sibuit.'] 

scabellum=V. scauel. 

scala=W. 3. yscawl, pi. ysgolyon. 

(?) scandere=W. 3. yscynnu ; W. 2. 
eskenho, eskynho (scanderit). 

schola=[sco/, gen. scide~\, V. scol; V. 
scolheic=W. 3. yscolheic (scholasti- 
cus), pi. W. 2. escoleycyon, pi. ysco- 
higyon; W. 2. escolectaut (status 

sciens : V. skientoc ; P. slcentyll, shyntyll 
(sapiens) ; Arm. squient (spiritus, in- 
telhgentia) ; V. diskient (insipiens), 
gaan ascient (" energuminus"). 

scribere=scribend ; V. scriuit, scriuen 
(scriptura), scriuiniat (scriptor) ; P. 
screfe (scribere). 

scrinium=sc?'i'n m. 

scripulus=W. 1. scribl; O. Gaedh. 
lethscripul (dimidio scripulo). 

scutella=V. scudel, P. scudell (discus, 

[sebum, V. suif,"W. 3. swyf, Arm. soav.~\ 
securus=P. sur. 

[senior, seinser.~] 
sensus=sens, dat. pi. siansib. 
sepelire=Arm. sebeliaf. [sepultura, sa- 

(? septimana = sechtmaine). [V. sei- 



EbeVs Celtic Studies. 

[septuaginta, septien.] 
sernionarius=Arm. sarmoner. 
[serus, W. hicyr.~] 

sextarius— W. 1. hestaur, pi. hestoriou, 
3. hestaicr f., (the h in the loan word 
is remarkable). 
[Med. Lat. sicera, V. sicer.] 
signum=[seV], Arm. sin. 
[situla, [Mid. Ir. sitheuL] 
solariuni=[Mid, Ir. soiler,] V. soler. 
solitarius=Arm. soliter. 
[Med. Lat. solta,,"W. swllt^v. sou.'] 
(somniari) Fr, songer=Arni. soingcf 

[sophista, Mid. Ir. soifist.~] 
(sors) Fr. sorte=Arm. sceurt, i.e. sort 

soutenir (Fr.)=Arm. soutenet (susten- 

spatium=W. 3. 
sperare: Arm. esper (spes). 
[spina, Mid. Ir. spin.] 
[spiraculum, spirucuL] 
s~pmtvis=s pirut, V. spirit, Arm. speret. 
spoliare— W. 3. yspeilaw ; dispeilaio 

(denudare, gladium). 
[spongia, sponge.'] 
[sponsa, Mid. Ir. pusta, W. pivys, 

stabuluin=[W. 1. stebill, pi.], W. 3. 
ystubyl. — V. steuel, W. 2. estauell, 3. 
ystauell f. (triclinium, cubiculum) 
appears to belong also to this place ; 
but compare also Fr. estaminet. 

[stannum, Mid. Ir. stanamhaiL] 
status=Arm. stat. 
stendardo (Romance), W. 3. yston- 

dard f. 
stimulus=W. 1. sumpl. 
stola=V. stol. 
stragulum=V. strail (tapeta), strati 

elester (matta). 
strata=W. 2. strut istrut, 3. ystrut 

(vallis aperta, planities). 
stratura (M. Lat )=\_sruthur], "W. 1. 

strotur. (stravi=W. 1. strouis?). 
strigilis=V. streil. 
superlativus = superluit superlit, pi. 


synodus=[Mid. Ir. senudh], V. sened. 
[taberna, Mid. Ir. tuibherne.] 
[tabes, tdm.] 
ta\entu.m=tullund (facultas, ingenium, 

Fr. talent). 
tardare=Arm. turdomp (tardemus), 

tardet (tardate). 
[tellus, telluir, gen. tellrach.] 

ternpero="W. 1. temperum (condio), 
templuni==:te??i/w/, Arm. P. tempel. [W. 

3. teml] 
tempt are=P. t empty e. 
(?) tendere=W. 3. tynnu ; Arm. emten- 

net (se recipere), teniff (pergam) ; P. 

tensons (tetenderunt). 
terminus=P. termyn (terminus,tempus). 
[tertia (hora), teirt.] 
testis=?es£, V. tist, Arm. test, pi. W. 2. 

testion; iestimomxim=teslimin, V. tis- 

tuin, P. tustunny; W.2. testu (testari). 
[theca, tiuch.] 
[theoria, teoir.] 
[thesis, teis.] 
thronus=Arm. iron. 
thus : tus-hstar (turibulum). 
[Titan, Mid. Ir. titul] 
titulus=^frta/ titol, ace. pi. titlu. 
[Fr. tonneau, V. tonnel.] 
tomeamentum (M. Lat.)=W. 3. twrnei- 

t&rqxLQS=muin-torc, W. 3. torch. 
torta=[for*Q, W. 1. 3. torth (panis). 
tTd.ct\is=[tracht],W. 3 trueth (sabulum 

maris), V. trait (arena), 
[totus, tot-muel gl. Totus Calvus.] 
traditio (Fr. trahison)=P. treuson. 
tribunus : trebun-suide (tribunal). 
trinitas==*n»ic?di£, [W. 1. trintuut], Arm. 

[tripus, TV. 3. tribedd, V. tribet.] 
tristis: [TV. 1. trist, P. trest], TV. 3. 

tristit tristyt tristweh (tristitia), tristan 

(tristem esse). 
tructa=V. trud. 
[truncus, TV. 3. truck, V. trech.] 
(?) tuba=gen. tuib. 
[tunica, tuinech.] 
(?) turba=TV. 3. ticryftwrwf. 
turris=[ft«V], W. tier, m., pi. 3. tyreu, 

tyroed, V. tur. 
[tympanum, timpun.] 
ultima=?<ft, ace. u.Ut. 
uncia = i*H<7ae unga (see above: ato- 

unguere, unctare=P. untye. 
[unicornis V. uncorn.] 
ursus=V. ors. 
(?) va.gma.=[Jaigen], TV. 3. gwuin, V. 

guein, P. goyn. 
velum=/?«/ (velamen), [V. guih] 
venenum=TV. 3. gwennwyn ; gwenivynic 

(venenosus), V. guenoinreiat (vene- 

[versatile ? Mid. Ir. fersuid, TV. 3. 

versus=/ers, gen.jersujerso. 
? Yerus=fir, V. Arm. guir, TV. 3. 

gwir, P. gwyr.- ? 

On the Position of the Celtic. 107 

vetus: Jetar-Iaice fetarlice fetarlicce viperae— W. 3. pi. gwiberot. 

(vetustas). virtus=P. vertu ; 0. Gaedh. gen. ferto 
[? vidua = fedb, V. guedeu, W. 3. ferte, nom. pi. forte, ace. firtu (virtus, 

gweddw.j prodigium). 

[vigil, figil] [visio,^?*.] 

villani=W. 2. pi. byleynyeyt. vitium=Arm. vice, (Fr. "vice), 

[vinea, fine.'] [vocula, Jocu1.~] 
yiau.m=finjinn, K. guin. 

To these are to be added a number of French words in the 
Armoric. Even from this list, which would, of course, be 
greatly enlarged if we were to include the more modern words, 
and in which, no doubt, many old loan-words are certainly only 
accidentally wanting, we can see what numerous borrowings took 
place already in ancient times, from the Latin, Middle Latin, 
and Romance. And even though the borrowing be doubtful 
in the case of some words (certainly not many), nevertheless 
the majority of the apparently exclusive correspondences of 
Celtic and Latin have been thereby removed. 

The Lathi has taken other words from the Gaulish, partly 
already in the classical period ; and later also from the British 
(as covinus) ; in any case, however, their number is not very 
great ; in regard to some of them there exist, too, doubts, which 
at present we are unable to solve. Those are especially impor- 
tant, which, although taken at a late period, have nevertheless 
passed into the Romance languages (as vertragus =It., veltro, 
etc., from this again V. guilter, molossus) ; but for our present 
object we may here fitly pass over these. 

§. 2. Glossarial affinities of the Celtic and Classic languages. 

The Celtic has about the following words and roots, in com- 
mon exclusively with one or both of the classical languages (or at 
all events, with such peculiarities of form or meaning as only 
recur in them). 



aer diar m., W. 3. aioyr=arjp, aura (or ardch=L&t. arduus (Gr. 6p96g appears 
derived from aer perhaps itself bor- to have been FopOog), not a loan- 
rowed?) word as the Gaulish Ar duenna 

ag (root): atomaig (impellit me)=Lat. (" heights", Gehenna " ridges") shows. 

adigit; — ayto, ago. as (a, es)=l£, ex. 

ailighn (muto)==-aXKaaa(x). V. auhel, Arm. auel {aura), W. 3. 

*ainm, W. 3. enw, P. (h)anow, Arm. awel (flatus). — V. anauhel (procella), 

hanu=6vofia (in the form). =ae\\a ? 

[aft=Lat. artus.] V. W. 3. auon f. (flumen), pi. W. 3. 

W. 3. alarch m., V. elerhc=L&t. olor? auonyd=Lat. amnis? 

anim, V. Arm. enef=. Lat. anima [ball=(paXX6g.'] 

(Zeitschr. VI. 213). V. her, W. 3. bereu (veru)— 0. Gaedh. 

[arba (read arva~), "W. 3. erw, V. erv, berach, birdae ("verutus") — = Lat. 

ere?«=Lat. arvumJ] veru (CJmbr. berva, berus ?). 


Ebel's Celtic Studies. 

? bethe C'buxus"), [W.bedeu], V. bed- 
even (" populus'') = betida betidla ? 
(according to Pliny, Gaulish). 

*bou; O. Gaedh. bdchaill, V. bv.gel 
(bubulcus, pastor), TV. 1. &ou^ (sta- 
bulum) ; also W, 1 and 2. V. buck 
(vacca) ?=/3o{/c, bos (in the form ; 
the other languages preserve the 

V. Arm.-6/-ec^, TV". 3. breich=lrachium, 

[bnide ' x e\\ov< ^ =b a di us, Pr. bai bay.'] 

can (root) : W. 3. lcanu, P. ca«e (canere), 
TV. 3. datkanu (recitare, revelare), 
O. Gaedh. forchunforcanimforchanim 
(praecipio), foircthe (eruditas), jor- 
cited forcetal (doctrina), forcitl(a)id 
forcetlid (praeceptor), tercital (va- 
ticinium), doaurchanaim (sagio), 
cet la /(/(cantor), also eel (augurium) 84 
and gen. ciud (instrumenti musici)? ; 
— "V. cheroot (cantor), canores (can- 
trix), — Lat. cano. 

TV. cann, V. can (albus)=Lat. Cand- 
idas (cf. the loan-words under can- 

car (root) (widely ramified in both 
languages, no doubt also the source 
of the, so far as I know, exclusively 
French form cherir) — Lat. cams? 

TV. 3. karw, V. caruu=Lat. cervus (if 
it be not borrowed? — the O.H.G. 
hiruz shows another suffix). 

cathir (civitas), K. cair caer (oppidum) 
=Lat. castrum ? 85 (compare as to the 
phonetic relations sethar siur, TV. 3. 
chwior=Goik. svistar). 

claideb, TV. 2. cledif, 3. cledyf, Arm. 
clezef, P. clethe (fundamental form 
* cladibas)=^Lat. gladius. 

cldi, TV. 2. cloeu pl.=Lat. clavi. [a loan- 
word ?] 

endm (ps)=Kvr)firi ? 

corp, K. corf==Lat. corpus, [a loan- 
word ?] 

cos (pes)=Lat. coxa ? [cosia ?] 

cretim (rel. cretes crettes creites, pi. 
cretite), TV. 3. cret (fides), Arm. cridif 
=Lat. credo, (see Stokes Beitr. I. 
458). [loan word.] 

[cruitr TV. 1. V.], V. croider (perhaps 
also 0. Gaedh. criathar, gl. cere- 
brum ?)=Lat. cribrum. 

TV. 3. civydaw, P. coihe=Lat. cadere ? 
(in the form rather=cec?ere). 

ail (tergum)=Lat. cuius. 

di. TV. 1. Arm. di, P. the, TV. 3. y= 
Lat. de. 

du- (do-~)—()vQ-, Skr. dus-. 

[TV. 3. ffer=ucpvp6v.] 

[/i=Lat. virus, ioe, Skr. visha.] 

V. gurah, TV. 3. gwrach (anus) — -/pave.? 

[ibim, Lat. bibo, Vedic pibdmi.] 

inis gen. tnse, TV. 2. mis, 3. y^s f., 
Arm. enes=Lat. insula ? (if perhaps 
this be a diminutive formation, not 
as Pott would make it=th'a\og). 

itir etir etar, P. intre yntre, Arm. entre 
(foreign to the TVelsh)=Lat. inter. 
[Skr. antar.] 

ith gen. etha (frumentum), V. yd 
seges) : Lat. ador (interchange be- 
tween d and t in ithim likewise). 

V. yorch (caprea)= Z,6pl,, dop% ? 

TV. 3. keissaw (instead of *kessiaiv, 
*kassiau) scarcely=Lat. quaerere? 

(Gaulish XayKia Diod. Sic. V. 30. pro- 
bably an erroneous supposition, other- 
vr\se=laneea, Aoy%?7). 

[lacht], V. lait (he), TV. 1. lafever 
(" lacocula"), TV. 2. laethauc (lac 
praebens)=Lat. lac. 

*le'ic (sine), leicci (sinit): general, but 
in the form exactly=linquit. 

liac, TV. 3. llech f. ; lapis, XiOoe, Xaiy£? 

lobur [ W. llwfr] (infirmus) : Lat. labor, 
labo, labes? 

loth gen. loithe (palus, coenum), TV. 3. 
lludedic (coenosusj ; Gaul. Luteva, 
Lutetia — Lat. lutum. 

matin, V. metin, TV. 3. yr meitin (mane) 
— Lat. matutinus (borroAved ?). 

[mil], V. mel=mel, p.e\i. 

midiur-sa (puto), TV. 3. medivl medol 
(cogitatio), medylyaw (cogitare) — 
p.klo\xai, meditor. 

TV. 3. mynyd, V. menit, P. meneth= 
Lat. mons ; emineo ? 

naue (gen.), noe=navis, vavg. [Skr. 
ndv, ndu.] 


nert, K. nerth (virtus) ; dvr\p, Osc. Umbr. 
ner, Sab. Lat. nerio, Nero. 

TV. 2. 3. oet. (aetas), 3. hoedel hoedyl 
(vita), oetaivc oedawc (aetate provec- 
tus)=Lat. aetas? (v could have 
dropped out as well in Celtic as in 
Latin, but compare also the still un- 
explained ui, oe in the verbal substan- 

[saiget] , TV. 3 saeth, P. se^=Lat. sagitta 
[borrowed ?] 

84 [Compare with this rather Old Korse heill, omen.] 

85 [Castrum is probably for cad-trum, and cannot be connected with caihair, 
gen. cathrach— an Old Celtic *cateirax, cfr. cataracton.] 

On the Position of the Celtic. 


sai=Lat. sagum. 

*saillim=a\\ofiai, salio (Goth, salta a 

different form). 
sarkail samal (similitudo)==-*sa?wa/t; 

amail amal (=dat. loc. *samaii) W. 

2. mal, P. avel, Arm. euel (ut) ; cos- 
mail cosmuil cosmil (=* consamali),\7 . 

3. kyffelyp kyffelyb (consimilis) — Lat. 

W. 3. sarff—Ski. sarpa, Lat. serpens 

sciath (0. Arm. scoit-, scoe£-)=Lat. 

scutum ? (the vowel is different). 
secA (praeter, extra, supra), K. hep heb 

(sine)=^sec?./s, t/caf . 
W. 2. Ae^'c, V. Ae%e/u=Lat. sa&r. 


su- [W. fo-]=Skr. sw-, Gr. ev. 

tar, W. 2. tries, 3. e?ros, P. c/ris drys 
(Arm. dreist)=Lsit. trans, Umbr. trdf. 

W. 1. tarater, 2. taradyr=Taperpov, 

tarvos (Gaulish), O. Gaedb. tarb, W. 2. 
taru,F. tarow=taurus, ravpog (conse- 
quently to be separated from Slav. 
turu, O. Norse \>ior, Goth, stiwr). 

tlr, K. tir (terra) nearest affinity— Osc. 
teeriim (possibly Lat. terra). 

[iian, W. 3. oen, V. oin=agnus, for 
avignus, " ewe-born" ?] 

urde, W. 3. gwyrcl, V. <7«iV* = Lat. 

Jdith=L&t. vates (borrowed ?). 

3. Glossarial affinities of the Celtic, Classic, Teutonic, and 
Lito- Slavonian languages. 

Others may no doubt be placed side by side with Latin ones, 
but are not the less Teutonic, Slavonian, and Lithuanian. The 
following occur more or less generally, for instance: — 


actus ocus (vicinus), comacus (vicinus), 
comaicsiu f. (vicinia), W. 3. agos, P. 
ogas (vicinus), W. 2. kauaens, 3. 
kyfagos (propinquitas, vicinitas), V. 
carogos (affinis, consanguineus) — 
first in the Greek lyyvg, ayxh but 
also Lat. angustus, O.H.G. angi; 
Slav, qza jaza, azu vazu (vinculum), 
Lith. anksztas ankszta (N.H.G. enge). 
The conjunction acus ocus ocuis (et) 
appears to be a dat. loc, as it has the 
power of aspirating. From the same 
root comes octe ochte (necessitudo), 
compare Lat. angor, angustia, Slav. 
jeza (morbus). Interchange be- 
tween cc and ng occurs elsewhere 
likewise, e. g. in cumacc, cumang, 

<3i7e,K. all=alius, aXKog, Goth, alisalja- 
(O.H.G. ali- in some few words, 
among which may be mentioned 
elithiotic, as was already observed by 
Graff — W. 2. alldut, pi. alltudion); 
in this form (with I) it is wanting in 
the Slav, and Lith. 

ainm (see supra)=Goth. namo, Slav. 
ime, Prussian emnes. 

dis dis, Gen. disa aisso, dissa desa (aetas), 
W. 1. ois (seculum), 3. oes (vita), V. 
huis (seculum), P. oys (aetas) ; nearest 
affinity=Skr. dyus, but then also aiwv, 
aevum, Goth, aivs, O.H.G. ev:a; is 
wanting in the Slav, and Lith. 

ar (root) (arare) ; general in all Euro- 
pean languages. 

athir (K. tat, like Gr. T8Tra)=pater, 
Trarrjp, Goth, fadar ; is wanting in the 
Slav, and Lith., which again differ 
from each other. 

ben, ban (mulier), V. benen (sponsa), 
benenrid (femina), benenuat (matrona), 
P. benyn (mulier), pi beny?ias=yvvrj, 
Boeot. (3dva, Slav, zena, Goth, qvens 
gvino, O.H G. chona ; is wanting in 
the Latin and Lithuanian (however 
there is Prus. *gannci). 

bar, ber (root), (Jerre) general. 

brdthair brdihir, W. 1. braut, 3. brawt pi. 
brodyr, V. brand broder, Arm. pi. 
breuder=f rater, (pprjrrjp (Zeitschr. 
VII. 436), Goth, brojpar, Slav, bratru 
bratu, Lith. brolis. 

bou (see supra) = O.H.G. chuo, Slav. 
govedo, Lettish gows. 

bin be'o (vivus), bethu beothu (vita), beod 
(vivus), biad (victus, esca), beoigidh 
(vivificat) ; W. 3. byw, Arm. beo, P. 
beu (vivus), V. bin (vita), W. 3. bywyt 
m., Arm. buez buhez, P. bewnas be?"- 
nans (vita), Arm. benaj (vivam), P. 
6ewe(vivere) ; vivus, (Siog, Goth, qvius, 
Lith. gyvas, Slav, zivu, etc. 

cride n. =icapdi.a, cor, Goth, hairto, 
Lith. szirdis, Slav, srudice. 

camm (curvus, obliquus), dat. pi. cam- 
maib, cammderc (strabo), camihuisil 


EbeVs Celtic Studies. 

(casus obliqui), W. 2. 3. Arm. V. 
cam. (curvus), V. camhinsic (injus- 
tus), Gaul. Camba, Cambodunum, 
Mopnca[i(3ri ; Gr. /ca/z7rrw, Lith. kam- 
pas a corner, kumpas crooked. 

[cnu, Lat. nux for cnux, hnot, Eng. nut.'] 

crmm f., Y.prif, W. 3. />r?// (vermis)= 
vermis, Goth, vaurms, Lith. (Jcirmis), 
kirmele', k\rminas, Slav. czruvx,czrwl 
(but czruminu), — Gr. sX/xivg?). 

cu, K. ci=Kvu>v, canis, Goth, hunds, 
Lith. szw (Slav, sw&a, sobakd). 

W. 3. cudyaw (abscondere, celare), P. 
cuthe, Arm. c«<ze£ (occultare)=/c6u0w, 
further Lat. cutis, O.H.G. hut f., 
hutta f. 

W. 3. keffyl (equus vilis)=Lat. caballus 
(icafiaXXrjg probably borrowed), Slav. 
kobyla, kobylica, koni, Lith. kumele, 
kumelukas, (kuinas, probably bor- 

[de'r], W. 1. dacr-lon (uvidus), pi. W. 3. 
dagreu, P. dagrow (lacrimae)=<5aKpt>, 
lacr-ima, Goth, tagr, Lith. aszara ; 
is wanting in Slav. 

daw (quercus), daurauch (quercetum), 
daurde dairde (quernus), derucc 
(glans), W. 3. V. dar, pi. deri, Sg. 
W. 3. derwen 86 (quercus); dopv, dpiig, 
Goth, triu, Slav, drevo (arbor), druva 
(ligna), Lith. derva ; is wanting in 
the Latin. [? Dr. Siegfried compares 
laurus from daurus, as lingua from 
dingua, lacrima from dacrima, 

dam (root) (in the Celtic, with a pecu- 
liar application of meaning) : fodai- 
mim-se (patior, tolero), W. 1 . guodeim- 
isauch (sustulistis), P. gotheff gothe- 
vell, Arm. gouzaf gouzaff (tolerare) 
=domo, da/jiaZto), Goth, timan, tamjan. 

det, K. dant m. (V. dans, pi. W. 3. 
danned)=dens, ddovg, Goth, tunbus, 
O. Norse tonn, O.H.G. zand zan, 
Lith. danth ; is wanting in Slav. 

dess, W. 2. dehou, 3. deheu, P. dyghow= 
Se£,i6g, dexter, Goth, taihsvs, Slav. 
desinu ; Lith. deszine (dextra). 

dia (dies), W. 2. diu dihu, 3. dyw along 
with dyd, V. det, P. dyth deth, pi. 
dethiow, Arm. deiz, pi. dizioa=La,t . 
dies, Slav, dim, Lith. dend; it is 
wanting with this meaning in Ger- 
man and Greek. 

dia (deus), W. 2. diu dyu dyuu dyhu 
duhu duo, 3. duw, V. dug, Arm. doe 
— W. 2. duyuaul (divinus)=c/e?;s, 
6s6g (?), Lith. devas, Lettish dews; 

is wanting in German and Slavonian. 
[But cf. O.N. tivar " gods".] 

dorus, W. 1. rc?r?<s], cfor, pi. 3. doreu, 
W. 3. G?nt?s, V. darat-=-Qvpa, fores, 
Goth, efowr dauro, Lith. durys, pi. 
Slav, e/w/7. 

ecA, K. ep=^equus, 'Linrog, O. Sax. eAw, 
O. Norse, idr; Lith, dssra (equa); 
is wanting in Slavonian. 

W. 3. V. Arm. elin, (ulna)=Goth. 
aleina, wXkvn, Lat. ulna; in the 
Lith. Slav, there is another suffix, 
where it is not wholly different. 

gaim-red, W. 1. gaem, 2. 3. gayaf, V. 
goyf, Arm, gouaff — W. 3. kynnhaeaf, 
V. kyniaj (auctuninus, i.e. forewinter) 
=/«ems, %twj/ %h/jwj/, Lith. zema, 
Slav, zwrca ; is wanting in German. 

1, gen (root Skr. jan) — in gigno (g)nas- 
cor, yiyvofiai, yevvdoj, Goth, kum, 
N.H.G. kind, Slav, zenti (gener), 
Lith. g\mti; appears in Celtic partly 
with g: nogigned (nascebatur), ro- 
gen(a)ir (na'tus est), dogentar gentar 
genthir (fiet), dogniu (facio), fogni 
(servit), gnim (factum), fogndm 
(servitus), congnam (contributio), 
gnethid (operarius), gein (ortus) Z. 
466, gen. geine Z. 1043, geinddae 
(genitalis), W. 3. Arm. ganet, P. 
genys (natus), W. 2. guneyr (fit), 3. 
givnaf(fsLcio, fa,ciam)=P. gwraffgraj, 
Arm. groaff graf gruif griff, etc. ; 
partly with c : cenel (natio, gens, 
genus), cenelae (genus), cenelach 
(generalis), cenaelugud (generatio)= 
W. 1. cenitol (generatio), cenitolaidou 
(natales), 2. kenedel, 3. kenedl kenedyl 
(genus), V. kinethel (generatio)= 

2 gen. (root — Skr. jna) — in yiyvwoTcw, 
(g)nosco, Goth, kan, Lith. zindu, Slav. 
znaja ; 0. Gaedh. adgen-sa adgeuin 
(cognosco), etargeiuin (noscit), gne 
(ratio), aitngne (sapiens, n. cognitio), 
irgnae, etargne, etarcne (cognitio), 
itargninim (sapio prudentia), nom- 
etargnigedar (me commemorat). 

ithim (mando), estar (edit), W. 3. ryt 
yssu (comesum esse) ; the primitive 
d in W. 2. heuedac (comessatio, 
epulae) ; the root ad is general. (The 
derivational ith (puis), "W. 1. iot 
"pulsum" appears to correspond to 
the Greek tidap). 

W. 3. ieuanc, pi. ieueinc, V. iouenc 
jouonc, P. yonk, sup. W. 2.3. ieuhaf; 
O. Gaedh. dclachdi (juvenilia), 

86 Erroneously explained in Zeitschr. VII. 211 ; — en is singulativ. 

On the Position of the Celtic. 


ticmil (tiro) =juvenis (juvencus), 
Goth, juggs compar. juhiza, Lith. 
jdunas, Slav, junii (Servian junak 
hero), (Gr. rjfin ?). 

W. 1. iou, V. ieu=jugum, Zvyog, Goth. 
juk, Slav, igo (i. e.- jtgo instead of 
jugo), Lith. jungas. 

lagait (parvitas), laigiu lugu (minor), 
lugimem (min\ma.m),W . 3. llei (minor) 
— skaaawv; Lat. lews, Slav, liguku, 
Lith. lengvas, Goth, leihts. 

Idn, W. 3. llawn, P. Arm. /e» Ze»tt (ple- 
nus), 0. Gaedh. lane Idine (impletio), 
lanma /r(impleti), linmaire (plenitudo) 
=plenus, Slav, pliinu, Lith. pilnas, 
Goth. /mZZs (i. e. fulns), the Greek 
has only the root, not the same deri- 
vatives; Goth, (fulljan) and Celtic 
have verbs derived from it: W. 3. 
llanw (implere), O. Gaedh. forldn 
(abundavit), rolin (implevit), comal- 
naclar (implet), linad (explere). 

il, compar. lia, etc. See Beitr., I. 

leth led (latus, dimidium), = lotus, 
rrXdrog, O.H.G. Mat plat. 

lethan, W. 1. lltan (latus) — tcXcltvq, 
Lith. platus, Goth, braids. 87 

ligim (lingo)=\a'Yw, lingo, Goth, laigo, 
Lith. laizau, Slav. liza. 

tnalg (root) ; Mod. Irish nieilg milk, old 
gloss do omalgg mulxi, Z. 71 ; every- 

man (root) : rommunus rommunus (scio, 
didici), domuinur-sa domoiniur clome- 
nar-sa (puto, spero), admuinur (volo), 
ni cuman Urn (nescio) and many deri- 
vatives ; W. 2. menoent (voluerint), 3. 
mynych mynnych (vis, voles), mynnir, 
(placet), P. mynny (vis), Arm. menaf 
mennaf minif (volo, posco, cogito), 
W. 3. gofyn, P. govynny (interrogare); 
memini, fikjxova, Goth. man, Slav. 
pameti (memoria) &c. 

mar (root) : marb (mortuus), W. 2. Arm. 
maru (mori) ; general (German and 
Greek only in derivatives). 

mdthir, suppressed in Kymric by the 
endearing word mam (=* mamma), 
preserved however in V. modereb, W. 
1. pi. modreped (matertera),=?nater, 
firirnp, O.H.G. muotar, Slav, mati; 
Lith. mote' (mulier). 

meddn, W. 2. meun, 3. mywn (medius, 
medium) : cognate words everywhere 
even though nowhere with this suffix. 

melim (niolo), damil-si (edis) ; general 
(Gr. nv\t] along with aXeio). 

mi (Beitr. 1. 461), [rnistae menstruus], 
W. 2. mis, 3. mys, V. mis=fxi]v, Ion. 
fieig, mensis, Lith. menu (instead of 
mines), Slav, meseci; deviating some- 
what Goth, mena (luna), meno)ps 
muir, K. mor (mare) ; general (Gr. 

masc (root — otherwise misc) : cummasc 
gen. cummisc (commutatio), commes- 
catar (miscentur), V. commisc, W. 
3. cymysc (mixtio), P. kemeskis, 
hemyskis (commixtio), Arm. kem- 
meski (misceo) ; O. Gaedh. cumsciget 
(mutant), rochumscigther (immuta- 
tum est), nicumscaichthi (non mutan- 
dum est), do not appear to belong 
to this root, because of conosciget 
(mutant), conroscaigissiu (summe- 
visti), conoscaige-siu (admoveto) ; 
misceo, /.liayoj, Lith. maiszyti, Slav. 
mesiti, O.H.G. misrjan. 
in niul u (in nubibus) = vt<p's\r), nebula, 
O.H.G. nibul; Slav. Lith. another 
suffix, and partly another significa- 
noct ; 0. Gaedh. innoc(h)t (hac nocte), 
[W. 1. henoid~], K. nos (nox) ; general. 
nu, [recte nua~] nue nuae, niiide (novus, 
novicius) ; general in forms which 
partly correspond to the Skr. nava, 
and partly navy a. 
V. ow=Lat. agnus, Ch. Slav, agnica, 

agnici jagnlcl, agne jagne. 
din den, K. wn=Lat. unus (O. Lat. 
oenos), Goth, ains, Lith. venas, Lett. 
wens. [Prus. ainsJ] 
V. palff.=palma, TraXdfin, Ang. Sax. 

O.H.G. folma. 
W. 3. V. rud (ruber) ; general. 
roth=L&t. rota, O.H.G. rad, Lith. 

salann, V. haloin halein (sal)=sa?, a\c, 

Slav, soli; Goth. salt, 
[suari], W. 3. V. Arm. hun=somnus, 
vttvoq, Slav, sunu (Lith. sdpnas, Lett. 
sapnis a dream), O. Norse svefn. 
smith (rivus, fiuvius, torrens), W. 2. 
frut f., 3. frwt., Y.frot=*Sk.Y. srotas ; 
Gr. psoj (apepio), Slav, struja struga, 
O.H.G. stroum, Thracian ^Tpv/aojv, 
Lith. sraume. 
V. hveger (socrus), hvigeren (socer)=r 
kicvpoQ eicvpa, socer socrus, Slav. 
svekru, svekry svekruvi, Goth, svaihra, 
W. 3. sych = siccus (O.H.G. biseh, 
bisihan Graff VI, 133?), Gr. aavKog 

87 [Much nearer is Old Norsey?afr=.English_y?a£.] 


EbeVs Celtic Studies. 

(aav\p.6g, aavcrapog, but also avxf-iog, 
avx/Mipog), Lith. sausas, Ch. Slav. 
sethar siur (also siar, fiar according to 
Stokes), TV. 8. c/acior, pi. cJucaer 
chwioryd, V. /m*r [recte Am'r] = 
Goth, svistar, Slav, sestra, Lith. sesfi', 
Lat. soror; it is wanting in Greek, 
unless perhaps sVapoc iralpog belongs 
to it. 

1 sak (root — to follow) : sechem (sequi), 
saigim (adeo), doseich (persequitur). 
saichdetu (consequential, sechimtid 
(sectator)=se9»or, t—ojiai, Goth. sa- 
Jean, sokjan, Lith. seku. 

2 sak (root — to say) : saigid (loqui, 
disputare), saiged (dicit), dosaig 
(dicit), saiges (g. dicit). insce (serrno), 
TV. 3. heb (inquit)=*'/?sece, ivv&tte, 
Lith. sakau, O.H.G. sag en. 

sad (root — to sit): insddaim (jacio), 
dordsat, doforsat (condidit, consti- 
tuit), adsaitis (residebant), sosad 
sossad (turris. positio), suide (sessio), 
suidiguth suidigud (positio) ; TV. 8. 
gor-sed-ua (sedes sublimis), P. set-va 
(sedes), settyas (posuit)=V£a;, sedeo, 
Goth, sitan, Slav, sesti, Lith. s'e'sti. 

W. 3. heul, V. heuul, P. houl (sol) ; per- 
haps, also, 0. Gaedh. soillse f. (lu- 
men) ?=Goth. sauil, Lith. sdule, Lat. 
sol (doubtful Gr. i']\iog=aFs\iog, it 
would be much better to consider it 
with Curtius=ows//. ? ); Slav, slumce 

[Gaulish 6'eno-magus], sen,K.hen=L&t. 
sen-ex (Gr. 'dvn), Lith. se'nas old, senis 
an old man, Goth, sin-eigs sinisto, 
O.H.G. siniscalc ; it is wanting in 

teg tech (domus), gen. idul-taigaz (fani), 
dat. taig (the fundamental form is, 
consequently, *tagi), W. 3. ty, pi. tei, 

Arm. ti ty, V. ti — from which tigerne, 

dat. tigerni (dominus), TV. 1. tigern; 

cf. Lat. ; Gr. rkyog, 0. 

Norse \ak % O.H.G. dach, Lith. stdgas 

(roof). [Ir. a-staig.'] 
temel m. (obscuritas), W. 3. tywyll (6b- 

scurus. obscuritas), V. tivulgou, P. 

tevolgorv (tenebrae)=Slav. tuna, Lith. 

tamsa, Lat. tenebrae, temere (blindly), 

O.H.G. demar (crepusculum) ; it is 

wanting in Greek. 
\tana~], W. 3. teneu (tenuis), P. teneiven 

(latus) ; tenuis, ravaog raw-, O.H.G. 

dunni, Slav. tiniku. 
tiiath, K. tut (populus)=Lith. Lett. 

Prus. tauta, Osc. tuvtu, Umbr. toto, 

Goth. ]>iuda ; it is wanting in Slav. 

and Gr. (as in Lat.). 
og, V. uy, W. 2. pi. uyeu=ovum, i66v, 

O.H.G. ei, Pol. jaje, Ch. Slav, ai-ce 

joke ; it is wanting in Lithuanian. 88 
fich (municipium. pagus)=i7c*/s, oIkoc, 

Goth, veins vehs, Slav, visi (prae- 

dium), Lith. vesz-pats lord, vese'ti to 

be a guest. 
fer, V. gur, W. 3. gwr (ground form 

* wra)==Lat. vir, Goth, vair, O.H.G. 

icer (weralt hominum aetas, seculum, 

generatio), Lith. vyras, Lett, icirs ; 

it is wanting in Slav, and Gr. 
fedb, V. guedeu=Got\i. viduvo, SI. 

vidova, Prus. widdewu, Lat. vidua ; 

it is wanting in Greek ; ijiOeog is 

scarcely connected. 
fescor, TV. 3. ucher, V. gurthuper, P. 

gicesper, Arm. gouspei'=vesper, sWt- 

pog, Lith. vdkaras, Slav, veczeru; it 

is wanting in German. 
TV. 1. gulan, V. gluan, Arm. g!oan= 

Goth, vulla, Lith. vilna, Slav, vluna, 

Lat. /a>?a ? ; Gr. anov is another 


To these are to be added tlie generally recurring roots Skr. 
as-, bhu, dhd (O. Gaedk denim (facio), Arm. doen doan, P. doyn 
(facere), and in tlie British compounds W. 3. bydaf, P. bethajf, 
Arm. bezajf, bizif), vid, cru (in all European tongues klu) 
and the numerals below 1000. If some of them are wanting 
in individual languages, it does not signify much for our pre- 
sent object, as here also we find everywhere agreements between 
the north and south. For example, daru [?] and the root 
sru are wanting in the Latin, vaskara, ghaima in the Teutonic, 

88 [The Teutonic words are scarcely connected with the Greek and Latin : 
O.H.G. ei, O.N. egg, A. Sax. dgg ; Crimean Gothic ada, point to original 
ADDIA, compare Skr. anda, egg.^\ 

On the Position of the Celtic. 


svastar, tamas, vidhavd in the Greek, dvja in the Lithuanian, 
akva, dacru, dant, sdna, and the root sak in the Slavonian. 
Even the absence of words from two languages (e.g. Lat. and Lith. 
gand, Tent, and Gr. diva, Slav, and Gr. sduala i tautd, vira, 
aina, Teut. and Slav, daiva) becomes for us of higher significa- 
tion, only when these are the two nearest related languages,* say- 
Latin and Greek, or Slavonian and Lithuanian. 

§. 4. Glossarial affinities of the Celtic, Classic, and Teutonic 


Of words which are wanting in the Lithuanian and Slavonian, 
the Celtic has the following in common with the two Classic 
languages and Teutonic : axle, athir, elin, palf nude ( ?) With 
the Greek and Teutonic it has, for example : 


dark (root — Skr. drc) : Arm. derch (as- 
pectus), 0. Gaedh. airdircc erdirc 
irdircc, pi. erdarcai (conspicuus= 
TrepiSt ptcijg ?), erdaircigidir (eonce- 
lebrat)=^lp/cw, O.H.G. zorht zoraht, 

* K. garan=ykpavog, O.H.G. chran-uh 
(in the form) in opposition to Lat. 
grus, Lith. gerv'e f, Slav, zeravll 
(* geravjas) m. 

lang (root) : loingtech (acceptus, gratus), 
fidang (tolerare), immefolhgai imme- 
folngai immolhgai (efficit), immejorling 
vmforling (efficit) — this form shows 
the composition, contrary to Zeuss 
756, notwithstanding arqfulsam (tole- 
remus) — indlung (findo), indlach (dis- 
ceptatio), cuimlengaithi (congredien- 
dum) ; cf. Xayxavuj ? O. H. G. ga- 

trag (root) : Gaul, ver-trogns ; O. Gaedh. 
traig, K. troit (pes) ; rpexoj, Goth. 

ban ben (root) : dofuibnimm (succido), 
eiirdibnet (perimunt), imdibenar (ab- 

sciditur), immeruidbed (circumcisus 
est), be'/nen pi. (vulnera, jilagae), [P. 
l(>iiij)itin~], tobe (decisio), nebthdbe nepJi- 
thdbe (praeputium), imdibe (circum- 
cisio), etardibe (iuterritus),6as(mors), 
batkach (moribundus) ; secondary root 
balm : epil (interit), atbela (morietur) 
=0ev in Qovog, ttsQvov, Tre<pii<JO[iai, 
ddvvijtpaTog, Goth, banja a wound, 
O.H.G. bana f. (homicidium), Initio 
m.(occisor), banon (quatere, exercere), 
0. Norse bana to kill, bani m. (occisor, 

(?) borg (borcc, borggde), P. burges (bur- 
gensis)=Goth. baurgs, also Gr. irvp- 
yog (cpovpicog)? — (may have been 
borrowed from the German). 

gen, dat. giun (os, oris) — compare xaivu), 
O.H.G. ginen ginon? ; the Latin 
hiare has different forms. 

[scath~\, V. scod (umbra)=Goth. ska- 
dus, Gr. (TKorog? 

sid, W. 3. hedwch (pax)=Goth. sidus, 
Gr. ZOog, rjOog? 

The following are Celtic, Latin (or Italic), Teutonic: 


Goth, alan, 

ad- 0. Gaedh. only in combinations, as 
already in Gaulish, Kymric ad- and 
at (difficult \ o be distinguished from 
aith=* ati, see Beitr. I. 3L2)=Lat. 
ad, Goth. at. [But also Lith. at.'] 

asil, K. esel (membrum)=Lat. ala, 
axilla, O.H.G. ahsala ? 

al (root): notatt (qui te alit), altram 

(nutritio) — Lat. ah, 

aljan, 0. Norse ala. 
\caecli\, V. cute (luscus, monophthal- 

mus):=Goth. haihs; Lat. caecus. 
W.3. crych (Gaul. Crixus?), O. Gaedh. 

crichaib (sulcis), W. 1 . clicked (ruga) 

=Lat. crisjms ; O.H.G. krus. 
Gaulish, Kapvov ri]v aaXitiyya, Hesych. 

114 EbeVs Celtic Studies. 

(kcl^vv%, Schol. II. a, 219). Kyrn. corn TV. 1. taguel. 2. taicel (silens), TV. 3. 

=Lat. cornu, Goth, haurn. leici (tacere), TV. 3. P. taw (tace)= 

gabor (caper), TV. 2. V. gauar, TV. 3. Lat. tacere, Goth. Ipahan ? 

gafar (capra) = Lat. caper, capra, red recht (lex), TV. 2. ra'M reyth reis 

O. N. Ao/r, Ang. Sax. hafer. (lex), Arm. reiz rez (rectus, Justus) 

gab (root), Kymr. cav (i. e. cabfi)= =Lat. rectus, Goth, raihts, O.H.G., 

Lat. capio, Goth, hafja. 0. Sax. relit n. 

li,W. 3. lliw, P. Zyzu (splendor, color, tenge, gen. sing. pi. te?igad=Qoth. 

gloria) — V. liuor (pictor), disliu (de- tuggo, Lat. lingua (dingua), although 

formis) — Lat. liveo livor lividus, with a different suffix in each lan- 

O.H.G. pit pliwes lead (Stokes). guage. 

nathir, V. nader=Jja.t. natrix, Goth, drog droch, TV. 3. drive, P. drok (malus), 

nadr, O.H.G. natra natara. (The W. 3. drycket (malitia), O. Gaedh. 

0. Gaedh. with its declension, stands drochgnim, ace. pi. drochgnimu 

as it were midway between Teutonic (malefactum), V. drocger (infamia), 

and Latin). drocgeriit (infamis), drochoberor 

nessa, superl. nesam=Osc. Umbr. ne- (maleficus) — cf. Lat. trux, O. H. G. 

simo, Goth, nehv-, already alluded to. triugan — Skr. root druh. [cf. 

[niae~\, V. noi=La,t. nepos, O.H.G. nefo, rpvx^ ?] 

A. Sax. nefa, and V. guins, P. givyns, Arm. guent (i.e. 

necht, V. wort=Lat. neptis, O.N., A. Kymr. *guini)=La.t. ventus, Goth. 

Sax., O.H.G. nift, O.H.G. niftila. vinds. [Skr. vdta.'] 

[nid], V. neid=L&t. nidus, A. Sax., caille (velamen) — cf. Lat. occulo, celo, 
O.H.G., nest. (The Slay, gnezdo is O.H.G. helan, heli f. (amictus, vela- 
obscure.) mentum), Goth, huljan, O.H.G. 

V. pisc=L&t. piscis, Goth, jisks. nulla (hiille) ; [caille is probably bor- 

[ri, gen.] rig, V. rwy=Lat. rex, Goth. rowed from pallium.'] 
reiks (O.H.G. rich). 

§. 5. Glossarial affinities of the Celtic, Teutonic, and Lito- 
Slavonian languages. 

Among the words the Celtic has in common with the Ten- 
tonic, Lithuanian, and Slavonian, besides those above quoted, 
which recur in Latin or Greek, we must no doubt remove 
many more which have come into all or several of these 
languages in the same way, by borrowing from the Latin, as for 
example: angelus z=I^ith. angelas, Ch. Slav, anigelu, O.H.G. 
angil, engil (Got. aggilus from the Greek), O. Gaedh. angel 
aingel; or apostolus ^lAth. apdsztalas, Ch. Slav, apostolu (Goth. 
apaustaulus) , O.H.G. postul, Arm., Com. apostol, O. Gaedh. 
apstal. Mutual borrowing among the other languages did not 
take place to anything like the same extent, and we run 
much less risk of mistaking the apparent relationships due to 
borrowing for primitive relationships, in this case, than in the 
comparison of the Celtic and Latin. The Cornish has borrowed 
the most from the Teutonic languages (especially from the 
English), like the Armoric from the Romance (French), next 
to them the Welsh ; the mutual influence between Gaedhelic and 
the Teutonic tongues may be considered as evenly balanced. 
The following are certainly borrowed: V. mesclen N.H.G. 
muschel, redior=:Eing. reader, herring Eng. herring, hot (capu- 
tium)=:Eng. hat, roche (fannus), streing (fibula) = Eng. string, 

On the Position of the Celtic. 


P. strek (radius sanguinis), strekis (plagae) = strike, V. strifor 
(contentiosus), P. stryff (contentio), strevye (altercari) = strive; 
W. 3. helym helm, iarll (conies) = O.N. iarl, A. Sax. tori, 
ysl(e)ipanu, to draw or tie (a bow), ysivein pi. ysiueinieit = O.N. 
swein (Eng. swain), ysmwg (vapour) = Engl, smoke. On the 
other hand, W. 2. 3. talu (solvere) and the German zahlen have, 
perhaps, come from a common source. 

The agreements of the Celtic with all three languages, or, at 
all events, with the Teutonic and Slavonian, at the same time, 
are certainly not exceedingly numerous, but for the most part 
all the more significant. Thus we again meet in Celtic with 
certainly a part, and very probably others, of those very words 
and forms which Schleicher has pointed out as exclusively 
common to the Teutonic and Slavonian. 



aball (malus), W. 2. aball (mali), abal- 
len, 2. 3. V. auallen f. malus), P. 
avell (pomum), W. 3. aual, pi. an rfeu 
aueleu (poma), Itwylbre/nd (uisdi)-— A. 
Sax. dppel, Frisian appel, O.H.G. 
aphid aphol, 0. Norse epli n. (malum); 
Lith. obel\s f. (malus), o6«7as=Lett. 
dbols (malum); S\a,v.jabluko jabluka 
(pomum), ablani ablorii jablani (ma- 

crocann, crocenn (receptaculum), W. 3. 
crochann (vas, olla), V. crogen (con- 
cha) — cf. O.H.G. kruog (lagena, 
amphora) ; Ch. Slav, kruczagu (vas 
fictile), hrucziminica a drinking house, 
krucziviniku host ; Lith. karczama 
a drinking house (N.H.G. krug) — 
Lett, krogs borrowed. — ? 

crauett f. W. 2. (pala fornacea), V. 
graidor (sculptor) — Goth, graham,, 
SI. greba (fodio), grobu (sepulcrum), 
Lith. grabas; the root is also no 
doubt Gr. (ypa^w), but with this sig- 
nification it is, according to Schlei- 
cher, Teutonic and Slavonian. 

doddUm (fundo), foddli (distinguit),jfott- 
drodd (qui id di visit), fodlaidi (divi- 
dendus), fodad fodd (divisio) — also 
dil (gratus), ddiu, ddem ,- ddes (pro- 
prius, certus, fidelis)? — W. 1. didaul 
(expers)=Goth. dadjan (dividere), 
SI. deldi, Lith. dalyti, Lett, dallit to 
share, Prus. dellieis imper. share with; 
Goth, dads, Lith. dalis f . a share. 

W. 1. drogn (coetus), drog (f actionem), 
i.e. no doubt drogg drong — cf. the 
perhaps Gaulish, drungus (a troop) 
with its un-Latin anlaut [_dr^ — Goth. 

dr'mqan to perform military service, 
gadrauhts a warrior, O.H.G. truhtin 
(dominus), truhtinc (paranymphus), 
O. Norse droit f., pi. drottir a troop, 
servants, drottinn lord, drdttning 
queen ; Lith. draiigas, dradgalas a 
companion, partner, SI. drugu (socius, 
alter, amicus), Lett, draudse {i.e. 
* draugid) a community. 

du do, K. do dy di y=Goih. du, A. Sax. 
to, O.H.G. za zi zuo, Slav, do, Lith, 
da-, is wanting in 0. Norse. 

Ir. droighean, Welsh draen has been 
compared by Grimm (Gesch. d. d. 
Spr. 1028) with Slav, trunu, Goth. 
\>aurnu& ; this comparison is, how- 
ever, only right if a guttural be sup- 
posed to have fallen out in the Skr. 
trna, Goth, and Slav., so that Lat. 
truncus and Gr. rtpxvog might also be 
connected therewith. O. Gaedh. drai- 
gen (" pirus") and V. drain (spina), pi. 
V. Arm. drein, P. dreyn, are found in 
Zeuss along with 0. Gaedh. driss 
(vepres), dristenach (dumetum), W. 
3. dryssien f. (frutex) — ? 

V. er (aquila)=Goth. ara, Lith. en's 
(Beitr. I. 234), erelis, Lett, erglis, 
Slav, orilu. 

[_ged],V. guit(a,uca),i.e. *gidd (anser)= 
N.H.G. genter,A.S&x.gandra, O.H.G. 
ganzo, Pliny ganta; also Lith. gan- 
dras (a stork) ? 

viang mace (root) (already spoken of in 
the Zeitschr. VI. 238 in the significa- 
tion augere, also in existence in deri- 
vatives mar (magnus), mace (filius) — 
the Goth, mag (possum)=Sl. mogq,, 


Ebel's Celtic Studies. 

Lith. mdku moke'ti (to be able, to un- 
derstand, to count, pay), etc., are spe- 
cially represented by O. Gaedh. cu- 
maing cumuing (valet), cumang (po- 
testas.posse). cumacc (potens), cumacht 
cumacht(a)e n. (potentia)=TV. 3. ky- 
foeth kyuoeth (potestas), 0. Gaedh. 
curaachtach (potens), comp. cumacht- 
chu (potior)=TV. 3. kyuoethawc (po- 
tens), V. chefuidoc (" ornnipotens"). 
Esj)ecially the Lithuanian tenuis 
agrees in a wonderful manner with 
the Celtic forms. 

menicc menic, TV. 3. mynych.'P. menough 
(frequens)=Goth. manags, Slav. 
mnogii (multus) ; 0. Gaedh. meince 
(abundantia)=Goth. managei, N. H.G. 
menge; mencain (penus). 

nocJit-chenn (nudus capite), P. noyth, 
Arm. noaz = Goth. naqva)ps, O. 
Norse naktr (nakinri), O.H.G. nachat ; 
Slav, nagu, Lith. nit gas. — The Lat. 
nudus is a different form ; it is want- 
ing in Gr. 

TV. 3. priawt, V. gur priot (sponsus), 
Arm. priet (maritus)=0.H.G. t /hW;7 
fridil, M.H.G. vriedel (amasius), also 
used for the husband) — Lith. pre- 
telius, SI. prijateli (amicus). Either 
priawt is to be compared with brawt 
(frater). therefore almost exactly— 
Q.H.Gr. friudil, &c. (with /for r), or a 
participle (amatus), to which the 
"Welsh per. pass. TV. 3. -at, -et, -it, 
-icyt, -aict), and Arm. part, (-et) ac- 
curately agree : in the latter case the 
adj. priaict (proprius), — from whence 
also TV. 2. ampriodaur (dou possi- 
dens), — represents the Homeric (piKoc, 
and N.H.G. "werth" (cf. yny priawt 
person, in (his) proper person). It is 
in any case one of the most interest- 

ing agreements between the Celtic, 
German, Slavonian, and Lithua- 

sil (semen), TV. 3. hen (serere), hewyt 
(satum est) — connects itself to a root 
form, which, according to Schleicher, 
is exclusively Germano - Slavonian : 
Goth, saian, SI. sejati, Lith. se'ti, 
sekla, se'mens. 

The root rdd, no doubt general, but 
in certain significations only Celtic, 
Teutonic. Litho- Slavonian (cf. Beitr. 
I. 426 seq.) 

snechti (nives) — the root is general 
(fundamental form *snigh). but the 
s has only been preserved in the 
northern languages: Lith. snegas, 
Slav, snegu, Goth, shaivs (=*snaigas, 
*snaigvas~) ; in the Gr. ayavvupoQ 
there is still a trace (=*aya<jvixFoc) ; 
in the Lat. nix nivis (=*nihvis, nigvis) 
it has wholly vanished. 

flaith f., gen. flatha flatho (imperium), 
fla(i)themnacht f. (gloria, dignitas, 
gradus), flaithemnas (gloria), TV. 3. 
gwlat (regio), pi. gwladoed gwledyd, 
V. gulat (patria), P. gwlas (terra) — 
fundamental form *vlati with the 
same transposition of the medial to 
the tenuis as in ithim (edo) — V. vu- 
ludoc (dives), TV. 1. guletic (potens), 
3. gwledic (imperans, princeps)— 
Goth, valdan, SI. vladiti vlasti vlada 
(imperare), Lith. valdau valdyti, 
Lett, ivaldit (N. H. G. walten, to 

TV. 3. gwerth (pretium), gwerthawr 
(pretiosus), P. gwerthe (vendere), 
gorthye (venerari) — Goth. vair)>s, 
Lith. vertas, Pruss. werts (the latter 
was perhaps borrowed, as the Polish 
wart certainly was?). 

§. 6. Glossarial affinities of the Celtic and Teutonic. 

Tlie correspondences with, the Teutonic are most numerous ; 
some of them are no doubt the result of borrowing, while 
in the case of others, the relation is not clear; many, how- 
ever, give no occasion for such a supposition. Compare for 
instance : 


agathar (timet), aichthi (timendus)= 
Goth. 6g, ogan. 

aithirge ithirge (poenitentia), aidrech 
(poenitens), taidirge i. e. do-aithirge 
(misericordia), P. eddrek, edrege, 
poenitentia)=Goth. idreiga. 

arbae orpe n.=Goth. arbi; orpam m., 
pi. horpamin=GrOth. arbja, comarpe 
=gaarbja; comarbus (cohereditas) : 
nomerpimm (trado me, confido), no- 
birpaid (tradite vos, confidite), nachi- 
berpidsi (ne conf.), roerbad (com- 

On the Position of the Celtic. 


rnissum est) pi. roairptha ; innarbar 
(abigitur, reoiovetur), arenindarbe 
(ut abigat), nackimrindarpai-se (quod 
non me repulit), arnachitrindarpi- 
ther (ne sis exheredatus) represents 
exactly N.G.H. enterben. 

bnga (contentiones), bagim (glorior), 
bdgul (praeda) — O.H.G. bdgan biag 
(contendere, objurgare), bdgen (con- 
tendere), bdga f. (contentio), O. 
Norse baga (obstare, resistere), baeg- 
jash (vexare, molestare), bdgi m. 
(difficultas), bdgr (molestus). 

bldil biail budil,W. 1. bahell, 2. bin/all (se- 
curis), 1. laubael (handbill)=O.H.G. 
bihalpihal bigil pig'd, M.H.G. bit — stdl 
unexplained in both languages. 

bolg bole (uter), Gaul. bulga==Goth. 
balgs, O. H. G. pale, O. Norse bdgr 
(follis, uter). 

borg=Goih. baurgs (see supra). 

(?) V. boch=O.H.G. boch pocfi, 0. 
Norse bokki, A. Sax. bucca — cf. O. 
Gaedh. cuilennbocc (" cynyps' 1 ) — ■ 
borrowed from the German, accord- 
ing to Grimm. 

W. 3. bwa (arcus) — 0. Gaedh. fidbocc 
(arcus ligneus)=0. Norse bogi, A. 
Sax. boga, O.H.G. bogo poco. 

W. 3. bwrd bord m., pi. byrdeu (mensa) 
=Goth. bawd a board, fotubaurd 
a footstool, 0. Norse bor% n. a board, 
table, ship, 0. H. G. bort borti borto 
m. (ora, navis, mensa). 

W. 3. blodeu, V. blodon (flos)=O.H.G. 
bluot f , M.H.G. bluotm. f., pi. bl'dete. 
— the Lat.yZos has a different suffix. 

brden (pluvia)=Goth. rign. The root 
also in /fyg^w, Lat. rigo, the special 
word-formation only recurring in the 

budid f. (victoria, bradium), buide boide, 
Z. 611 gratiae), biddech (gratus, con- 
tentus), ho-buidnib (copiis); W. 1. 3. 
Arm. bud (bradium, victoria, for- 
tuna), W. 1. budicaid, 3. budugawl 
(victoriosus, felix), W. 1, bodin (tur- 
ma), pi. bodiniou, 3. by din f. — the 
root is the same as in Goth, ana- 
biudan, faurbiudan (jubere, mandare), 
the fundamental signification was pro- 
bably to announce=Skr. bodhaydmi 
(denuntiare) ; cf. N. H.G. aufgebot 
with bodin. — The signification is dif- 
ferent in Slavo-Lithuanian. 

(?) W. bad m., pi. 3. badeu (scaphae)— 
cf. 0. Gaedh. bddud (naufragium)=: 
boot, not High German, 0. Norse 
bdtr, A. Sax. bat — borrowed from the 
Celtic, according to Grimm. 

cath, K. cat (pugna) — Gaul. Caturiges, 
Catuslogi — O.H.G. hadu (only in 
names), A. Sax. heaZo, M.H.G., 
N.H.G. hader. 

(?) "W. 1. can; 2. car, Gaul, carrus 
(Caesar). — O.H.G. Icarra garra 
charra f., O. Norse kerra (appears to 
have come into German through 

W. 3. craff (firmus), P. cryff, cref 
(fortis, gravis), Arm. cref creff (fir- 
mus, tenax), criff (fortis), craf (ava- 
rus), W. 3. kyngryfet (aqua fortis), 
craffu (f-utiter incedere), creff t (ars) 
— cf. O.H.G. chraft (not in Tatian), 
A. Sax. craft, 0. Norse Icreftr, and 
hramph. — ? 

cruim (curvus)=O.H.G. chrump, A. 
Sax. crumb (remoter and doubtful 
Lat. eurvus, Lith. kreivas, Slov. krivu). 

"W. 2. cussan, V. cussin (osculum) — 0. 
Norse, A. Sax. coss, O.H.G. elms. 

dorche f. pi. (tenebrae)=A. Sax. deorc, 
Eng. dark, O.H.G. tarch, O. Norse 
dockr (obscurus). 

dun (arx), AY. 2. 3. din (castellum)= 
0. Norse, 0. Sax. A. Sax. tun, O.H.G. 
zun, Engl, town (on the names of 
places see Beitr. II., part 1). 

gabul (furca, patibulum) = O.H.G. 

guide Q)ilo praeditus), Gaul. Gaesati, 
gaesum^O.H.Q. get; A. Sax. gar. 

gell (pignus)=Goth. gild (tributum)? 
(see giall). 

V. gJiel (sanguisuga)=O.H.G. egala, 

W. 3. gerihi (virga), V. garthou (sti- 
mulus) may, no doubt, be compared 
with the O.H.G. gartja (switch), 
but the Goth. gazds=O.B..Gr. gart, 
N.H.G. gerte (goad, switch, whip), 
points to a borrowing into Celtic 
from the Teutonic. 

giall (obses) (gell (pignus) Z. 64, see ' 
supra), V. guistel (obses), W. 3. 
gwystyl (obses, pignus), P. gustle 
(spondere), Arm. goestlas (spopondit) 
=O.H.G. gisalf 

Q)glass glas (glaucus), Arm. glisi (livor, 
aegritudo) — 0. Norse, O.H.G. glas, 
A. Sax. glas (vitrum). 

V. grou (arena) — O. Norse griot (la- 
pides, saxa), A. Sax. gr'eot (scobs), 
O.H.G. grioz (glarea). 

(V. hos (ocrea), W. 3. hos(s)an, pi. 
hossaneu (braccae) = O.H.G. hosa 
(caliga), A. Sax. hos (calcaneum), 
hosa (caligae). Evidently borrowed, 
but by whom ? 


EbeTs Celtic Studies. 

(W. 2. hucc (sus), V. hoch (porcus)= 
Engl, hog ; the latter appears to have 
been borrowed from the Celtic (A=s, 
therefore related to sus). According 
to Grimm, it was the Celtic which 
borrowed from the German, N.H.G. 
haksch (verres) — ?) 

iarn (gen. hiairn. Inc. Sg.), W. 3. heyrn, 
V. hoirn (0. Arm. haiar?i-,hoiarn-)= 
Goth, eisarn, O.H.G. isarn, 0. Norse 
isarn iarn. 

[eo], V. hiuen (taxus)=O.H.G. iwa f. ; 
A. Sax. iv, 0. Norse yrm. (cf. Zacher 
das Goth. Alph. p. 10. seq.) 

\_ldr], W. 1. laur, 3. llawr (solum), V. 
lor, P. ler tear (pavimentum, solum) 
— with the dropping of p=-flur, 
M.H.G. vluor? (Grimm 307 also 
compares A. Sax. flor, Engl, floor). 

V. Arm. lag at, P. lag as, W. 3. llygat 
(oculus) — A. Sax. locian, O.H.G. 
luogen, N.H.G. lug en — Skr. root lax? 

land: dat. isind- ithlaind (in area), W. 
3. lann (area, ecclesia), 0. Arm. lann, 
Ital. Fr. Provencal landa, lande= 
Goth. land. 

V. loven (pecliculus)==C>.H.G. A. Sax. 
O.N. lus (?) 

Mm (saltus), "W. 1. lammam (saho), 
lemenic (salax), W. 3. llemhidyd (sal- 
tator)=M.H.G. limpfen to limp (Jam 
N.H.G. lahm, Engl, lame) ? Thence 
also W. llamp=Qoih lamb (the hop- 
ping) ? 

loathar (pellis)=O.H.G. ledar, O.N. 
ledr ; A. Sax. le%er (funis)? — The 
meaning would answer, yet the 
Gaedh oa and the O.H.G. e differ. 

marc, K. march (Gaul. ace. [iapicav)= 
O.H.G. marach, f. meriha, M.H.G. 
march (marc). 

mi- (is wanting as a prefix in Kymrie) 
=Goth. missa, N.H.G. mis-. To this 
is to be added the comp. messa (pe- 
jor); further W. 3. gormes f., pi. 
gormesseu gormessoed (miseria, afflic- 
tio, infortunium). 

mong, W. mwng, pi. W. 1. mogou 
(read moggou i.e. mongou)=mdhne, 
O.H.G. mana, once manha, M.H.G. 
man (?). 

mucc, W. 3. moch (sus), according to 
Grimm. N.H.G. mucke (?). 

W. 3. ychen pi., O. Arm. oAe?i=Goth. 
auhsans (the Latin vacca deviates). 

o's, uas, uch, Corn, ugh, Arm. us (supra), 
gen. uasal, K. uchell (altus) — cf. 
Gaul. Uxellodunum, Brit. o'v^tWov, 
ovt,t\\a — Goth, auhuma, auhumists 
(supremus). The Picenian Auximum 

has a different meaning (Zeitschl 
III. 248). 

lobar, Kymr. lavar (loqui), 0. Gaedh. 
amlabar, V. aflauar (mutus), mab 
aflauar (infans)=N.H.G. plappern 
(blappen, blappern) ? — Bopp com- 
pares Skr. lap, the I appears however 
to be old. 

run, K. n'n=Goth. runa, 0. H. G. run 

sam (sol), W. 1. ham, 2, 3. V. haf, Arm. 
Aq^"(aestas) — 0. N., 0. H. G. sumar, 
A. Sax. sumor sumer. — Also Goth. 
sunna, sunno, A. Sax., O.N. sunna, 
O.H.G. sunna sumna ? Pictet and Leo 
Meyer (Zeitschr. IV.) have explained 

scoloca (servi [scholastici?]), banscala 
(servae) — Goth, skalks ? 

seol sdol (velum, carbasus), W. 1. huil, 
V. guil (velum)=0. N. segl, 0. H. G. 
segal, A. Sax. segel. 

set (via), dat. se'it, pi. seuit seuit ; se'tche 
(uxor), dat. seitchi (properly a female 
fellow-traveller, Gefdhrtin; a word 
from the nomadic time?) ; W. 1. hint, 
Arm. hent (via), thence V. camhinsic 
(injustus), eunhinsic (Justus)— Goth. 
sin)>s ; 0. H. G. sind m. 

slid m. pi. (ostreae), Sg. slice (lanx)= 
O. H. G. snecco (Umax), A. Sax. sne- 
gel (limax, cochlea, testudo), 0. N. 
snigil (limax), more especially M.N.L. 
slecke (Umax). 

snathe m. (filum), dat. sndthiu; V. snod 
(vitta), V. W. 3. snoden (filum), W. 

3. ysnoden (vitta) cf. 0. N. snara 

(laqueus), 0. H. G. snuor f. (filum) 
from the same root. 

[such and] W. 1 . suh (vomer) according 
to Haupt in Z.=rO.H.G. sech(?). 

tre, tri, Kym. ^rui=Goth. \>airh (Beitr. 

idle, K. o//=Goth. alls. 

(h)uathath (Ji)uathad huathad hothad 
(singularis, singularitas), gen. uathid 
hodid, dat. othud uathuth, ace. hua- 
thath ; uaithed (singularis, solus, soli- 
tarius); f. ace. pi. huathati (singu- 
lares), dat. pi. uathataib (t=thth) ; 
dthatnat (pauculus) — of one stem with 
da (minor) from Skr. ava: but also 
comparable with Goth. au\>eis (de- 
sertus) N.H.G. oc?e=Skr. "avatya. 
[? Lat. pau-cus]. 

fen (plaustrum), Brit. Belg. covinus= 
0. N. vagn, O.H.G. wagan, A. Sax. 
vdgen. (The Greek and Slavonian 
have different suffixes). 

fladnisse (testimonium)=O.H.G. giwiz- 

Oji the Position of the Celtic. 119 

nesi f., n., A Sax. gevitnesse, folcaim folcaimm (humecto, lavo), W 

gevitnes, Engl, witness. 2. 3. golchi, Arm. guelchi, P. go/hi/ 

.fid n., K. guid, Gaul. vidu-=0. N. vfiSr (lavare) — A. Sax. volcen, O. Sax. 

m.,0. Sax. vidu,A. Sax. w*rfw, O.H.G. wolcan, O.H.G. icolchan (nubes) as 

t«*ta n. (Beitr. I., L60), with an equal moist or moistening ? 

change of meaning, thence, for ex- TV". 3. gicyllt, V. guilt, P. gwi/Is=Goth. 

ample, V. colviden (corylus), with vilpeis. 
the singulative suffix. 

§. 7. Glossarial affinities of the Celtic and Lito- Slavonian. 

The exclusive agreements between- the Celtic and the Lito- 
Slavonian are very much less numerous. To these belong, for 
example : — 


Gaedh. K. bran (corvus)=Slav. vronu, TV. Arm. merch (filia, puella), V. moroin, 

Lith. vdrnas (corvus), vdrna (cornix). TV. 3. morwyn (puella) [O. Ir. moru] 

TV. 3. Arm. gatlaf, P. gallof (possum, perhaps=Lith. merga, mergele? 

potero)=Lith galiii gale'ti. TV. 1. 2. melin, 3 melyn (flavus, lividus), 

Gaedh. nem, K. nef, SI. nebo n., Lett. f. 1. melen, pi. 1. milinon, 3. melynyou, 

debbes f. with the signification hea- V. milin (fulvus, flavus)=Lith. me'- 

ven (contrary to Lat., Gr., Lith., and lynas blue? according to Diefenbach 

Germ.). (Beitr. I. 483) from M. Lat. melinus 

c«jVe £ (accusatio, nota, culpa), cairigud = jjurjkivog, in this case, however, 

m. (reprehensio), [TV. 1. cared, gl. ne- we ought to expect Kymr. muffin, 

quitiae,] TV. 3. kergd m. (reprehen- moilin. 

sio) — Ch. Slav, karati (rixari), Lith. (?) glun, TV. 3. Arm. glin m. (genu) — 

Jcoravdti (punire). perhaps=Sl. koleno (genu), Lith. 

Gaedh. cruim in the form=Lith. Jcirmis kulnis, heel, kelys knee? (g instead of 

(the Lat., Gr., and Germ, have lost kin gabor, gabimm also), 
the &). 

Finally, the Celtic also is of course not wanting in words 
which heretofore have not been found in any primitively related 
tongues, or, at least, in any European language. Of the first 
kind is, for example, tene, K. tan, in contradistinction to the 
Skr. agni, Lat. ignis, Lith. ngnis, SI. ogni, as well as to the 
Greek 7rup, Umbr. pir, O.H.G. fur: among the special agree- 
ments with the Sanskrit, the similar nomenclature of the points of 
the compass (Z. 67. 566) is particularly remarkable. 

All these glossarial agreements and deviations would of course, 
taken by themselves, prove very little, as we find even between the 
most nearly related idioms, striking differences, such as between 
Slav, and Lith. in the case of the name of God, between Lat. and 
Umbr. in the appellation of fire. Where, however, the same or 
nearly related words recur in great numbers, there we have at 
least every inducement to further investigate whether special 
agreements may not be found in the grammar also, and in this 
expectation we are rarely disappointed. Among the words and 
forms quoted in the preceding pages (and I believe I have been 
perfectly impartial in their selection), there recur exclusively 


120 EheVs Celtic Studies. 

about fifty undoubtedly in the Teutonic tongues, not quite 
forty certainly in the Latin ; if to these we add about twenty 
which certainly recur in Latin and Teutonic, about a dozen in 
Latin and Greek, at least as many in Teutonic and Lito- 
Slavonian, it follows that the degree of relationship between 
the Celtic and Teutonic on the one hand, and the Celtic and 
the Latin on the other, is pretty nearly the same, with however 
some preponderance to the side of the Teutonic, which is still 
further somewhat strengthened by the few Teutonic-Greek agree- 
ments. The Lithuanian and Slavonian on the one hand, and the 
Greek on the other, are decidedly further removed as regards 
glossarial resemblances, being as compared with each other about 
equal. With the Celtic they are chiefly connected by the Teutonic 
and the Italic tongues. The Celtic prepositions also show that a 
similar proportion is to be expected in the grammar ; among them, 
for instance, ad is again found in Teutonic and Latin only, di and 
tar only in Latin, ire only in Teutonic, and du in Teutonic and 
Slavonian. 89 The prefixes du- and su-, which otherwise are every- 
where wanting, lead nearer to the Greek, while the privative cm- 
is again found in Greek, Teutonic, and Latin (the SI. Lith. u- in, 
for example, iibagas, ubogu, appears to correspond rather to the 
Skr. ava-y 

§. 8. Phonological affinities; — Vocalismus. 

In Phonology, the principles according to which we might 
judge of an earlier or a later separation of tongues, are as yet 
by no means finally established, and agreements between unre- 
lated languages, and differences between the nearest related ones 
here present themselves often so strikingly, that we should avoid 
deciding about their relationships according to such data. 
Thus, for example, the treatment of the mutes in O. Gaedh. 
agrees in the most wonderful way with that in the Hebrew 

Cfe *^> n -> ^ even T: : instead of IE?), while the Polish 
wholly departs from the Slavonian rules (as in ivilkz=1uith. 
vilkas, in opposition to O. Slav, vluku). I think that a geo- 
graphy of sounds is chiefly wanting to arrive at a conclusion as 
to how far the phonetic laws of languages are affected by phy- 
sical, genealogical, or social influences; 90 in this the vocalismus 

89 \Du occurs in composition in 0. Lat. : in-dw-perator, in-c/w-pedio.] 

90 [I am glad to find that so competent a philologist as Ebel has come upon this 
idea of a geography of sounds, which, so far as I am aware, I was the first to put 
forward, though crudely, in Vol. II. of the Atlantis. If such a man as Dr. Ebel 
were to turn his attention to this subject, the foundation of an important branch 
of science might be laid. Briicke's attempt to classify all the articulate sounds 
which could possibly be produced by the tongue ( Grundziige der Physiologie und 
Systematic der Sprachlaute. Wien, 1856), affords a basis to begin upon, for if we 

On the Position of the Celtic. 121 

as well as the consonantismus, and the relations of both to one 
another, should be taken into account. The above-mentioned 
phonetic similarity of the Gaedhelic and Hebrew, for instance, 
appears to be due to similar physical conditions; the spora- 
dically occurring one of the Polish with the Lithuanian to social 
(historical) circumstances. The agreement already pointed out 
by Lottner of the Goth, mikils with the Gr. and Lat. /uiyag 
magnus, in contrast to the Skr. mahat, appears to point to a closer 
relationship between the European tongues ; so in like manner the 
Goth, daur with Gr. and Lat. dvpa, fores, in contrast to Skr. 
dvdra. One of the most important points in connection with, 
and most conclusive evidence of, earlier or later separation of 
the individual languages, namely, the elementary develope- 
ment of the vocalismus, can be followed out with clearness 
unfortunately only in a single language, the Gothic. The 
Gothic triad of the, short vowels a, i, u (as in Skr. and O. Per- 
sian), speaks unanswerably for a proportionably early separation 
of the Teutonic from the other European tongues, at a time 
when none of the then united languages had developed an e and 
6; in like manner the Lithuanian must have separated from the 
Slavonian before the latter had developed an 6; the Lithuanian 
from the Lettish before the long a was changed into 6. The 
Latin and Greek, on the other hand, admit of the assumption 
of e and 6 before their separation. The sign no doubt only, and 
not the sound of o, was wanting to the older Umbrian and the 
Oscan. The Latin and Greek afford a marked contrast to the 
Teutonic in the circumstance, that perhaps everywhere in them, 
certainly at least as the rule, the a has been changed into z, only 
through e; in the Latin also through o into u; in Teutonic, on 
the other hand, it is the reverse, a being changed into e through 
i, and into o through u. The Celtic takes in this respect so far 
a middle place, inasmuch as a direct passage of a into o (and e) 
cannot be denied already in Gaulish nominatives like ^eyo/uapog, 
and accusatives as vefj.r}Tov, as also in the (primitively long) 
Old Gaedhelic genitive endings -o (I. 177, 180); it places itself, 
however, by the side of the Teutonic by the circumstance that 
in both riving branches u has passed into o, i into e (not o, e 
into u, i), and just as in Teutonic partly by breaking (fer: 
*firas = O.H.G wolf: Goth, vulfs), partly by simple weakening 
(Arm. ed: Welsh and Cornish yd = O.N. son: Goth, sunus ; 
compare also Slav, snocha, deni for older snucha dint), and the 
u and i appear here also for Sanskrit a, without the middle stages 

knew all possible sounds, and could classify them, we would merely have to 
determine in what part of the world each sound occurred. I hope to return to 
this subject at another time. — W.K.S.] 

10 b 

122 EbeVs Celtic Studies. 

o and e, as in Teutonic and Slavonian: coic, Kym. pimp = Qoth.. 
fimf =$kr. panca ; Gaul, dula (irajunrtBovXa probably Graecised) 
= Skr. data; duine Kymr. dyn, den (homo), perhaps from a root 
dan = Sav in Svtjtoc? Decide, therefore, as we may regarding 
the interesting agreement spoken of in Beitr. I. 163, of the 
Celtic and Teutonic in breaking and umlaut, whether we recog- 
nize herein with Lottner (Zeitschr. VII. 27. cf. Schleicher KS1. 
Formenlehre p. 11) a certain family likeness, or in consequence 
of its later origin, leave it with Schleicher (Beitr. I. 442) unno- 
ticed, the direct passage of a into o and e (ocht, echznocto, eguus) 
should not, at all events, be looked upon as a proof of a closer 
relationship to the Latin, especially as it also occurs in Slavonian, 
the o of which nevertheless was evidently originated only after its 
separation from the Lithuanian. Vowel-changes analogous to 
those in Teutonic and Slavonian are besides also found in the 
Celtic roots: guidimm (precor), where ui is umlaut from u, along 
with ro-gdd (rogavi), foddli (distinguit), along with fo-ro-dil 
(di visit), Idnad, along with linad (complere), brdth, along with 
breth (judicium). I will not, however, lay much stress upon 
all these agreements, in consequence of the uncertainty which 
still generally prevails in such questions. But in the diphthongal 
system the Celtic comes decidedly nearest to the Teutonic, and 
at least much nearer to the Lito- Slavonian than to the Latin or 
Greek. The Teutonic starts from four diphthongs: ai, ei, au, in, 
and after all the changes has returned in New High German to 
four : ai, ei, au, eu. The Celtic most distinctly leads back to four 
diphthongs : ai, oi, au, iu. The Lito- Slavonian appears also to 
have had only four diphthongs before it divided, to which the 
' Slav, e, i, va, it, and the Lith. ai, ei, and e (both = Prus. ei) au, 
u, point back ; the Lith. ui and Slav, y appear to be of later 
origin. In the Latin and Greek, on the contrary, six diphthongs 
evidently lie at the base of their system : ai, ei, oi, au, eu, ou. An 
interesting analogy, although of later origin, occurs between 
Gaedh. ia, ua, along with $, 6, O.H.G. ia, ua (ie, uo), along with 
Goth. 3, 6 (Grimm. Gesch. d. d. Spr. 844), and Lith. e, u, Slav. 
e, va for the guna diphthongs ; on the other hand the Kym. u 
= Gaedh. oi (oe) agrees with the Lat. it for the older oi (oe). 

§. 9. Phonological affinities; — Consonantismus. 

In its consonantismus the Celtic connects itself with the 
Lithuanian and Slavonian in this, that in its older phonetic 
stage it had no aspirate. The Kymric ch is throughout only a 
sharpening of the spirant h for s, as in Slavonian, and of similar 
origin, only that it has not attained the extension of the Slavonic 
ch; the Gaedhelic/is a hardening of the initial v, the Kymric/ 

On the Position of the Celtic. 123 

(jf ) is nowhere, as Zeuss thought, a primitive aspirate, but has 
arisen from s or belongs to loan-words (see Beitrage, II. 82), 
only the Gaulish / is still obscure. In this respect the Celtic 
stands in marked contrast to the Greek, with its three aspirates, 
somewhat less so to the Latin, which to be sure has no aspirates, 
but whose spirants / and h rest upon old aspirates. It deviates 
from the Teutonic inasmuch as the latter has preserved dia- 
lectically to the present day an aspirate th, and has also ch 
in the Frankish, but it agrees with it therein that, in both 
languages the aspirates which do occur are all hysterogens, and 
rest upon older tenues. Gaedhelic and Teutonic exhibit some 
agreement in this also, though it is of later origin, that the 
secondary aspirates have also frequently changed themselves 
into medials (or medial- aspirates) . The change of the old 
aspirates into medials is common to all European languages, 
in the Greek occasionally, chiefly after nasals ; in the Latin 
pretty regularly in inlaut; in the others almost without ex- 
ception ; here the Celtic and Teutonic agree best, because sibi- 
lants often take the place of old aspirates in the Slavonian 
and Lithuanian. The Celtic exhibits a remarkable approach to 
the Teutonic in the occasionally occurring hardening of the 
medials, as for instance in the root gen, where even the Gaulish 
affords the combinations Oppianicnos, Toutissicnos* 1 in tenge 
(along with Goth, tuggo, therefore, for * denge), in inlaut in 
itliim, ith, cumacc along with cnmang (here likewise in accord 
Lith. moku, as opposed to Slav, moga), rofetar (scio) along with 
Goth, vait =Ski\ veda. This looks almost like a beginning 
of the German provection of sounds; but on the other hand 
medials occur instead of tenues in gabor, Kymr. gavar=zlja.t. 
caper capra, Teut. * hafar, in Gaedh. gabdil = Cym. cavael, 
Lat. cape're, Teut. hafjan, in Gaul, ande-, Gaedh. ind- com- 
pared with Gr. avri, Goth, and-, with which the Lith. gelbetizz 
Goth, hilpan agrees. The Gaedhelic thickening of the n [rather 
nn~\ into nd in certain positions, Z. 54, is decidedly of later 
origin ; it has peculiar analogy to the Goth, hunds, N.H.G. je- 
inand, O.H.G. pliant ( = Fr. pan). Considering the ignorance 
which for the moment exists, as to how far phonetic relations 
may be taken as a measure of relationship, I have meanwhile 
thought it would be useful to also bring forward such agreements 
as are of demonstrably later origin, or which might appear in 
the present discussion of inconsiderable importance. 

§. 10. Affinities of word-formation. 

In word-formation, the suffix -Hon appears to be exclusively 

91 See Pictet's recently published Essai sur quelques inscriptions en langue 

124 EbeVs Celtic Studies. 

Italo-Celtic (the contraction to -tin only in Oscan, Umbrian, and 
Celtic), not much exclusively northern can be opposed to it; the 
use of -li as an infinitive suffix is akin no doubt to the Slavonian 
-lu in the participle ; it is confined, however, to the single gabdil 
and its compounds. Other suffixes aie generally, or pretty gene- 
rally diffused, such as -id in the feminine abstracts in e, -ti in the 
infinitives, 92 the latter seldomest in Latin. The following seem 
to have been borrowed: -aire -iW = Goth. -areis, Slav, -art, Lith. 
-orius (from the Latin -arius, which appears to have arisen 
from *-asius) ; and -doit = W. 2. -taut -daut, 3. -daivt, Arm. det 
(from Lat. tas), both chiefly in loan-words (likewise the Kymr. -es 
of the fern. = Romance issa from the Greek -taaa, and -ids m. 
= Romance -his from the Latin -ensis). The suffix-combination 
*-antat, [rather *-antdt\ in O. Gaedh. -atu, -etu (Z. 272) is quite 
peculiar to the Celtic. The Celtic word-formation, however, so 
far as it is known to us, bears a modern character like that of 
the Romance ; such a heaping-up of suffixes, as is the rule in the 
known Celtic languages, is a very rare occurrence in the Latin 
especially. The use of the suffixes has especially much more 
widely extended itself in composition; while, for instance, an 
aoTrAoc clvott\oq sufficed for a Greek, and an inermus, at most 
changed into inermis (instead of Hnermius ?) for a Roman, the 
O. Irish, like the Kymric, could scarcely attain in the Greek 
way (\6yog, aXoyog, aXoyia) to an amlabar (mutus) = V. 
aflauar, or W. 2. anuab (cltskvoq), but mostly had recourse to 
suffixes: cretem, ancretem, ancretmech (rr belief, unbelief, unbe- 
lieving). In general k especially has attained a much wider ex- 
tension than in the Classic languages : already in the Gaedhelic 
-ach plays as a determinative suffix a much more important part 
than in the Latin (senex) and Greek (yvvaiKog), and numerous 
forms such as apstallac(h)t, brithemnac(h)t may be opposed to 
the single senectus; but in the Welsh participles in -etie the -ic 
places itself completely by the side of the Slav, sladukii, etc. 
The Celtic agrees with the Teutonic, especially in the deri- 
vation of the ^verbs in -aigimm and -igur; while cumachtagimm, 
cumachtaigim still connects itself with cumachtach, like the 
N. H.G. bemdchtige with mdchtig; asmecnugur (eradico), nomis- 
ligur (humilio me) go quite as far beyond the limits as the 
N.H.G. peinige, reinige (In Graff. IV. 3, there are only three 
such verbs without adjectives: bimitnigon, chruzigon, tiligon). 

§. 11. Affinities of declension. 
As regards the declension, the circumstance which I have 

92 See ante, pp. 60, Gl. 

On the Position of the Celtic. 125 

already touched upon in the introduction, namely, that the 
so-called Pelasgic tongues only have feminine a-stems (-oc, -as), 
appears to me of importance ; the Celtic here agrees with the 
northern languages. Masculine a-stems, which, beside the Latin 
and Greek, occur also in the Slavonian and Lithuanian, appear 
to be just as foreign to the Celtic as to the Teutonic: com- 
pare, however, Stokes. 93 The Celtic has just as few feminine 
w-stems as the Lithuanian, 94 and at bottom also the Slavonian, 
whose -y (ui) is transformed into -uvi, -vi and -va (Schleicher 
K. Slav. Formenlehre, 214). On the other hand, it ap- 
proaches to the Classic languages at least nearer than the Teu- 
tonic and Lito- Slavonian in this respect, that it has preserved 
pure more consonantal stems ; it, however, again separates itself 
from them by the treatment of s-stems, and lastly the passage 
of vocalic stems into consonantal ones seems to be found in 
Europe exclusively in the Classic languages. The preservation 
of the ablative, if it were established, would certainly speak 
strongly for the connection of the Celtic with the Latin ; that has, 
however, as yet by no means been done, and least of all by forms 
like innurid** whose d could not possibly represent a primitive 
final ablative -d or -t. (In the opposite case the construction of 
prepositions with the dative would bring the Celtic close to Teu- 
tonic). I cannot lay the same weight as Schleicher does upon 
the preservation of the b in the dative plural ; the absence of any 
contraction in this case rather indeed places the Celtic nearest to 
Teutonic. But then it approaches the Greek and Latin by 
the total want of the peculiar pronominal declension, which no 
doubt, on the other hand, has left evident traces in the 
Umbrian pusme and esme, esmei. The agreement between the 
genitive singular and nominative plural of the masculine a-stems 
in Old Gaedhelic and Latin, appeared to me from the very 
first extremely remarkable; the deviation of the Oscan and 
Umbrian from the Latin in both cases on the one hand, and 
the reappearance of the fundamental form -ai in the nominative 
plural of the Lithuanian and Slavonian, as also the Greek, had, 
however, hindered me from drawing further conclusions from it, 
especially as I could never thoroughly convince myself of the 
correctness of Rosen's interpretation of the Latin genitive -i 
adopted by Bopp. The communication of the old locative 
forms by Stokes 96 now to be sure throws a new light upon 
this genitive also, and makes me more favourable to Bopp's 
view. To draw further conclusions from so wonderfully exclu- 
sive an agreement as that which the Latin exhibits to the 

93 Beitr. I. 464. 94 See ante, p. 58. 

95 Beitr. I. 454, 96 Beitr. I. 334. 

126 EbeTs Celtic Studies. 

Celtic, in opposition to its nearest relatives, remains, however, 
always attended with uncertainty, because the other agreements 
in the case-forms (dat. -iu -u, voc. -£, ace, pi. -us) recur every- 
where except in Teutonic. In the consonantal declension the 
gen. sing, -as, nom. pi. -is or -es, by the side of the Greek -oc> 
-£C (Old Lat. gen. -os -us), and in opposition to the Gothic -is, 
-as, bring the Celtic phonetically close to the "Pelasgic"; but 
similar points of contact are also found between very remotely 
related tongues. 

§. 12. Affinities of Gradation. 
In the gradation or comparison, the Greek isolates itself from 
the analogy of the other languages by its superlative suffix -tcitoq 
(simple -roc, and arog is also, except in numerals, foreign to the 
others), the Latin by its -issimus (=is + timus); the Celtic -am, 
-em (*-amas, *imas) likewise occur only sporadically elsewhere 
(in prepositional derivatives), its -imem nowhere. The Sanskrit, 
Greek, and Teutonic -ista, is wanting in the Latin and Celtic, 
and every proper superlative suffix in the Lithuanian and Sla- 
vonian (except remains like Lith. pirmas=Goth. frumd). The 
superlative forms in the Gaedhelic particle-composition iarm-, 
remi-, tairm-, tremi-, correspond to the Lithuanian pirm, Goth. 
fram (both used as prepositions and prefixes) ; com- before 
(vowels and) aspirated consonants, Z. 842, is no doubt a form 
of the same kind. I have already 97 mentioned a very significant 
analogy between the Celtic, Teutonic, and Slavonian, and at- 
tempted to explain the Celtic forms, — the double formation of 
the comparative in Old Gaedhelic -a and -iu (-u), Goth, -iza and 
oza, Slav, -ii (-iszi) and -ei: a similar relationship appears to 
exist in the Lithuanian between the comparative {-esnis, adv. 
-jails) and the superlative (-jdusias, adv. jdusei). 

§. 13. Affinities of the Pronouns. 

The Celtic differs from all its relatives in the pronoun in the 
giving up of the nominative singular of the first and second per- 
son ; for me, me (cf. Fr. moi) is either originally the accusative, 
or formed from the stem of the oblique cases, and til, tu, appears 
aspirated as a true vocative only in the combination athusu (o 
tu), otherwise it resists aspiration, and has accordingly been 
explained by Stokes as the accusative. But the pronoun of the 
third person exhibits in the noun e, si, ed, whose feminine we 
again find in the Kymr. hi, an extremely striking similarity with 
the Teutonic ; this exactly resembles the Gothic is, si, ita, and 
the retention of d in primary auslaut even appears to indicate a 

97 See ante, p. 91. 

On the Position of the Celtic. 127 

form *ita. The accusative feminine -se (as the t instead of d in 
inte, intesi, shows) and the accusative plural -su, -siu, (cf. intiu 
and the almost constant double r in airriu, erriu, erru, and 
constantly in etarru and form) correspond to the Old High 
German sia and sie, sip, siu; perhaps indeed the Old Latin forms 
like sum, sos, may likewise be here compared, but not in the 
nom. sing. fern. We only find in the Sanskrit forms corres- 
ponding to the genitive di, d pi. ah; 9s so likewise to the dative 
plural -aib, -ib — *abis [rather -abo] (compare doib, doib, doaib- 
sem along with 2. duib, duibsi, foraib,forib, along with 2. foirib 
fuirib, indib is, on the contrary, common to 2. and 3). The 
pronoun ta, the use of which in its isolated form is foreign to the 
Latin, otherwise preserved everywhere, appears to be preserved 
in the dative uad, ood, f. uadi, plur. uadib, uaidib, the d of which 
cannot be easily explained otherwise, so likewise in indid. The 
pronoun ana, which is foreign to the Classic languages, and on 
the other hand is preserved pure in the Lithuanian ans, Slavonic 
onii, in the Gothic jains with a (hardly merely phonetic) addition, 
is evidently again found in the Celtic article, although it ap- 
pears there are in the Gaedhelic forms with a prefixed s also 
(from sa ?) 

§. 14. Affinities of Conjugation. 

But, most remarkable of all is the position of the Celtic with 
respect to all the cognate languages in the conjugation. Very 
peculiar combinations and new formations have occurred here, 
to such an extent that, for instance, the old ending of the first 
person singular present -u ( = Lat. -o, Gr. w, Lith. -u, Goth, -a, 
O.H.G. ~u, Slav, -a for primitive -*ami) has been preserved pure 
only in extremely few Old Irish forms: bin (sum), tdu (sum), 
dogniu (facio), deccu (video), tiagu (venio), tucu, tuccu (intelligo), 
roiccu (indigeo), togu (eligo), and is to be recognized in some 
others, at least by the umlaut, e. g. forchun (praecipio). Again, 
striking agreements with the Latin occur in the formation of the 
tenses and the passive. Notwithstanding these circumstances, a 
wonderful analogy with the Teutonic and Slavonian is found to 
exist, which points to a most special connection of these lan- 
guages, the result either of long continued unity, or of a very 
special relationship of the mind of the peoples. The Old Gaedh- 
elic paradigm completely connects itself with the Lithuanian in 
this respect, that the present and the praeterite have quite the 
same endings, not even deviating in the singular, as in the Greek ; 
compare, for instance — 

98 See ante, p. 73. 

128 EbeVs Celtic Studies, 






















The Kymric -st of the second person singular praet. has been 
looked upon as the more primitive form, and compared with the 
Latin -isti, although in the Celtic there is nothing in the plural 
analogous to Lat. -istis (Lottner, Zeitschr. VII. 41) ; that this 
explanation does not strictly apply, but rather that the Kym. -t, 
as Pictet" had already surmised, is, as in many other verbal 
forms, a relic of the pronoun (e.g. O. Ir. carim, cairimj, is 
shown by the corresponding O. Ir. deponential form: ru-ces- 
taigser (disputasti), which has no -t, while the third person ro- 
labrastar (locutus est) has preserved the -t (ill), which has 
frequently disappeared in the present, and always in the pre- 
terite. This seeming agreement may, however, be accidental, 
even unreal. The Kymric agrees more closely and certainly 
with the Slavonian, as Schleicher 100 remarked, in the combi- 
nation of the roots bliu-\-dJ\d; W. 3. bydaf= Ch. Slav, bqdq; 
but in a more general manner there may be also compared the 
Ch. Slav, idq, (eo) jadq. (ascendo), Goth, iddja (ivi) and the -da 
in German weak praeterites, -da- in the Lithuanian imperfect and 
present participle. This composition with -dhd extends farthest 
in Slavonian idq, and next to it in Welsh bydaf, bydwn, byd; 
even W. 3. oedwn (eram), the d of which is wanting in the pre- 
sent wyf, also appears to explain itself in the same way, and perhaps 
even the awd in the 3rd per. sing, praet. (Z. 504, frequentissima 
et omnibus verbis communis terminatio, ita ut in hodierna, lingua 
eadem (scripta -odd) sola pro hac persona in usu sit), though -awt 
in the passive, no doubt, also appears by the side of it. It is 
particularly remarkable that this -d likewise passes over into the 
root composition peculiar to Kymric (especially Welsh) so that 
for example in gwybydy (scis) three roots occur fused together, 
gwyd-\-bu-\-da, and in gwnathoed (fecerat), even as many as four, 
gwyn + aili + oe -\- da. All these agreements in particulars appear 
insignificant, however, compared to a pervading analogy in the 
Slavonian, Teutonic, and both branches of the Celtic, which has 
forced itself from the beginning, on me at least, as one of the 
strongest proofs of the correlation of these languages. 

As is well known, the Slavonian dialects mark the distinction 
between the imperfect and perfect, continuous and momentary 
action, which the Greek, Latin, and Romance languages express 

99 De l'affinite, etc. 150. 10 ° Beitr. I. 505. 

On the Position of the Celtic. 129 

by special tense-forms, by separate verbs, trie composition with 
prepositions playing therein a great part. Thus, for instance, 
almost the whole of stem verbs are imperfect in the Polish, but 
become perfect by composition. What appears strangest to a 
foreigner is, that the present is wanting in perfect verbs, because 
the form of the present has assumed a future signification ; but 
we again find the same phenomenon, because it is founded in 
the idea of the verb, in the Greek ei/ut, whose present has 
future, whose moods and imperfect, have aoristic signification. 
That this phenomenon does not, as it at first seems, stand 
isolated without any analogy in other languages, was shown by 
Grimm in his introduction to the translation of Wuk's Servian 
Grammar (1. seq.) and he expressly pointed to a similar dis- 
tinction in German (" starb" and " verstarb", " ich reise and 
" ich verreise morgen"), and also indicated that a still more accu- 
rate agreement with the Slavonic might be found in Old 
German. 101 Schleicher 102 has worked this out farther and more 
accurately, in the first instance only in relation to the future 
in the Gothic and Slavonian, glancing however at other forms 
which characterize the Gothic compositum as verbum perfectum. 
An extremely interesting point with regard to this has been 
overlooked, namely, the translation of the Greek part. aor. by 
the part, praes. of compound verbs: usstanda?ids avaaraq Math. 
c. ix. v. 9 ; gastandands araq Mark, c. x. v. 49 ; gahaasjands 
aicoiHjag c. x. v. 41. 47; ushlaupands av ai\ rjS/j (rag, afvairpands 
awofiaXcov v 50, andhafjands cnroKpiOeiG, v. 51, andbindandans 
Xticravrtg, c. xi. v. 2, gataujandan KaTEpyaaa/Lievov I. Corinth. 
c. v, v. 3; samcfy gagaggandam izvis avva\Oivrwv v/mwv, c. v. 
v. 4 (where Massmann, altogether wrongly, and entirely misun- 
derstanding this peculiarity, prints, contrary to the manuscript, 
gaggandam). The whole power to alter the sense here resides 
in the particle, which, when no other is present, is ga-. In New 
High German, such distinctions as also occur in the passage of 
Tatian, already quoted by Grimm : ihaz siu bdri, inti gibar (ut 
pareret, et peperit) have for the most part been obliterated, 
but sometimes petrified also: thus in the ge- of the part, praes., 
the prototype of which may likewise be found in Gothic, e. g. 
fulan gabundanana ttwAov SaSsjuivov, Mark, c. xi. v. 2. 4. 

What herein especially separates the Teutonic and Slavonian 
from other tongues which have something analogous, is the great 
force of the particle in composition, and we meet with a perfectly 

101 The verbs with a double theme in Greek and Sanskrit offer a somewhat 
analogous phenomenon, e.g., \afi(3dv(o imperfect, 'i\a(3ov perfect; compare also 
the future use of the conj. \dj3oj in Homer. 

102 Zeitschr. IV. 187 seq. 

130 EbeVs Celtic Studies. 

analogous order of tilings in the Celtic languages also. In the 
old languages, wherever another particle (O. Gaedh. m, W. 2. 
ed, 3. yd) has not effected its suppression, we also always find 
the idea of the perfect denoted by a particle, and as in Teutonic 
by ga-, by a special one: ru- (ro, ra, W. 2. P. re, W. 3. ry, 
Arm. ra), wherein I have already (infra, p. 163, with Stokes's 
concurrence, Beitr. I. 459) conjecturally traced the Sanskrit pra. 
This particle denotes exactly, as in Gothic and Slavonian, the 
perfect as well as the future, and, just as in German, its use in 
the modern language is limited and fixed for certain cases. The 
Celtic deviates in its grammatical form from the Teutonic in 
this, that its ru- remains before or (like the Greek augment) 
after other prepositions : ni roimdibed (non est circumcisus), im- 
meruidbed (circumcisus est), while the German ge- does not 
enter into true composition. The reason of this after-position 
of the ru- in Old Irish is obviously this, that here, as in the 
oldest Greek and Sanskrit, the prepositions remain in perpetual 
tmesis (sit venia verbo !) as the treatment of the so-called infixed 
pronouns shows: imm-itm-ru-idbed (circumcisus sum properly: 
me circumcisum est) ; forms like asrobrad (dictum est) therefore 
agree perfectly with German ones, such as ausgesprochen (in 
separable composition). The Kymric, which does not actually 
affix its pronouns after other prepositions, also does not put the 
ru- in the middle, but the Cornish and Armoric deviate therein 
from the Welsh, that the two former put the pronouns also before 
ru-, the latter allows them to follow. 

At the other side of the Channel we find this particle — Firstly 
before the praeteritum along with the usual sign of the tense : 
O. Ir. rorelus (manifestavi), W. 1. ro-gulipias (" olivavit"), 2. 
re-briuasei (vulneraverit), P. re-werthys (vendidi), re-wresse (fe- 
cerat), so also in the passive before the original participle, in 
order to denote the perfect: O. Gaedh. ro-noibad (sanctificatus 
est), P. re thyskas (instituti simt), W. 3. ry echewit (relicti sunt). 

Secondly, before the present and the future (like Gothic g'a- 
before the present participle), which are thereby changed into 
the perfect future exactum: O. Gaedh. ro-comalnither (com- 
pletum est), ro-ainmnichte (denominatum sit), arnachit-r-indar- 
pither (ne sis exheredatus), ro-beimmis (fuissemus), ra-n-glana 
(emundaverit se) ; the treatment of the infinitive in Welsh is 
extremely interesting in this respect: 2. e-re kafael (se invenisse, 
properly: suum invenisse), 3. ry-gaffel (accepisse), which accu- 
rately corresponds to that of the participle in Gothic. 

Thirdly, before present forms, especially the conjunctive and 
secondary present, which acquire thereby a future signification, 
as robia, robbia, ropia (erit), or, what is analogous to it, conjunc- 

On the Position of the Celtic. 131 

tive signification (cf. Gr. oirwg ttoii]gu, also firi Xaj3>?c, along with 
ju?7 Xa/xj3ave, like Latin ne dixeris), therefore goto- corro- conro- 
(ut), e.g. conrochra (ut amet), conrogbaid (ut sumatis), conrobam 
(ut shnus). We only find the second and third methods in 
Armoric, but here the custom of the language has gradually 
decided for the use in the conjunctive, which connects itself 
more especially with the third way. The Gothic also shows all 
three uses. 

The Gaedhelic has only so far passed beyond the limits of 
the Slavonian and German as to have also given a particle to 
the tenses of incomplete action, nu-, no- (explained by Stokes 
as the Skr. ami, Beitr. I. 470), only in simple verbs however, 
mostly also only to the secondary tenses, seldom to the primary 
present, and future. I will not even venture to make a surmise 
as to what the Kymric yd (W. 2. ed, P. y, Arm. ez), which 
occurs before all tense-forms, signifies, and what may be its 
origin ; 103 the Gaedhelic du (do), which we find instead of the 
ru (ro), does not differ, probably, from the preposition du; mu 
(mo) instead of nu (no) is obscure to me. 

The use of the particle before the future and for the future, 
was perhaps much more extensive in Celtic in ancient times, 
and has thus probably in part become the cause why the future 
has disappeared, in Gaedhelic in so many instances, in Kymric 
almost wholly; at all events, the Celtic is in most beautiful 
harmony with the Slavonian, and above all with the Gothic, 
as regards its use of the verbal particles. 

At least equally significant analogies of the Celtic to the Teu- 
tonic (and in a secondary degree to the Lito- Slavonian) as to 
the Italic (and further on to the Greek) have then everywhere 
presented themselves ; a kind of middle position will accordingly 
scarcely be denied to it. It appears, however, as if the pheno- 
mena which it has in common with the Teutonic were precisely 
those which chiefly indicate the intellectual life, the internal 
character of the language. In this category I include, besides 
the great extension of the composition with independent words, 
as well as with suffixes, the twofold formation of the degrees of 
comparison, and the importance of the verbal particles. 

In conclusion, it may be mentioned that a comparative syntax 
might bring to light many peculiar points of contact between 
the Celtic and Teutonic, such as the use of the infinitive with 
do, the government of the accusative by cen (sine) ; and that in 
general, the Celtic, so far as it is known to us, bears in its syntax 
so decidely modern a stamp that, to me at least, it is very difrl- 

103 [The Gaulish ate- ?] 

132 EbeVs Celtic Studies. 

cult to imagine its connection with the Latin to be so intimate 
as Schleicher does. In this respect, the Latin evidently bears 
the most antique stamp, the Greek a much more modern one 
(for instance, bj the freedom in the use of the infinitive and by 
the use of the article) ; again, the Lithuanian and Slavonian a 
much more antique one than the Teutonic ; but the most modern 
of all is the Celtic ; so that many things in the Romance lan- 
guages appear to rest upon Celtic peculiarity. Of this, perhaps, 
another time. 



§, 1. Necessity of establishing an organic Orthography; and 
great importance of a comparison of the Modern Irish forms 
for the purpose. 

SCHLEICHER has justly remarked, that an organic ortho- 
graphy is, above all things, necessary to enable us to get a 
right knowledge of the Old Irish language. This aim will, no 
doubt, be only to some extent satisfactorily attained when more 
extensive and more connected linguistic monuments shall be in 
our hands than we have at our service on the Continent, and 
when the editors will strive to attain a greater literal accuracy 
in their publication, than unfortunately appears to have been 
hitherto mostly done. Take a few examples in order to show 
how little, on the whole, one can trust to the literal accuracy of 
citations: — Zeuss quotes the same word from the same place 
three times differently spelled, 263 beisti, 1009 bessti, 1059 
bessti; O'Donovan gives the following from Cormac's Glossary in 
two different ways, 292 tibradaibh, 360 tipradaibh, so likewise 
151 carput, 252 carbat, as dative singular. Fortunately we see, 
at least in the first case (although we may remain in doubt as to 
the reading of the codex), by the Middle Irish oclit m-biasta, 
and na n-ocht m-biast (Visio Adamnani in O'Donovan 440, 441), 
as well as from W. 3. bwystuil, that the e is long, and conse- 
quently that bessti is wrong, and in both the other examples the 
tiprait of the Leabhar Breac (O'D. 249) and carpat in Cor- 
mac's Glossary (O'D. 3), as also the Latin loan-word carpentum, 
prove that the true O. Ir. form required two tenues p and i, 
which sunk to media? only in Middle and Modern Irish, — tobar 
(Keating in O'D. 394) and carbad. In tipra (or tipru?), gen. 
tiprat, an nt-stem (Stokes Beitr. I. 457), the^> appears, however, 
to have arisen from b, by means of the hardening action of an ori- 
ginal preceding mute, as in idpart, aedparthi, and in the examples 
in Zeuss 80, consequently ti- instead of tid- as taith-, taid- (Z. 
852) derived from doaith^—Ql tid-barid (offerte) Z. 253. 
But we have not everywhere at our disposal similar sources 

104 So likewise probably in timne n. (mandatum, prseceptum) from do- aith-mne 


136 EbeTs Celtic Studies. 

from which to obtain aid in determining the true old form, and 
where a new and unknown word presents itself to us, we are at 
present almost helpless. The necessity is then the more press- 
ing for Celtologists to use every available means for fixing the 
phonetic laws, and establishing an organic orthography. For 
this purpose the most important of all is the comparison of the 
Middle and Modern Irish forms, where this is possible ; the com- 
parison of the scanty remains of the Gaulish language, which 
are almost confined to proper names, and the Kymric dialects, are 
only of secondary importance, and last in order is that of the 
other Aryan languages. 

The Modern Irish is often so strangely disfigured, even in 
comparison with the Old Irish, and from want of literary cultiva- 
tion has (like vulgar languages generally) become so very irre- 
gular that a direct comparison of its words and forms with those 
of Sanskrit and the kindred languages would be very daring, 
and hazardous, in a still higher degree, for instance, than if we 
were to directly compare the New High German with the 
Sanskrit. Most of the errors in the first comparative investiga- 
tion of the Celtic by Pictet and Bopp were due to this cause, and 
it was only by the publication of old Irish forms in Zeuss' Gram- 
matica Celtica that a firm ground was gained and a solid founda- 
tion laid for Celtic philology ; everything correct that had been 
found before that time, we must consider as the especially lucky 
result of a wonderful divinatory faculty. 105 Who could, for in- 
stance, recognize the root gab (capere) in the imperatives fagh 
(find) fag (leave), tog (raise), the first of which has even a pre- 
sent faghaim, without such forms as the infinitives dyaghbdil, 
dydgbhdil, do tliogblidil, which have still preserved the ending 
consonants. The O. Ir. forms fagebtis (haberent, caperent) — 
together with fogbaidetu (usura) — , fodcbat (gl. deponant, i. e. 
relinquant) Z. 1072, foracab (reliquit), fotrdcbussa (reliqui te), 
fdcab (he left) Tir. in O'D. 437, lastly cotaucbat Z. 1072, and 
cotaocbat (attollunt se, surgunt) supply the explanation, and the 
Middle Irish faghbait, faghbat (they obtain, find) O'D. 241, fo- 
ghebha (thou wilt get) 242, and, on the other hand, fagbas, 
fagbus (he leaves), 155, tdgbhaidh (raise), 180, show the pas- 

(root man). Cf. taithminedar, taidminedor, taidmenader (significat, memorat) 
in Z. and Mid. Ir. timnais (he bids), in O'D. 155 ; damnae (Tirechan in O'D. 
436) appears=c?o-??2ne. 

105 Unfortunately M. Pictet has again lately (Beitrage, II. 84 sq.) trodden the 
same dangerous path. I cannot, according to what has been said above, recog- 
nize as conclusive, nor yet disprove, the examples which are there to prove the 
passage of p into/, so long as the corresponding older forms shall not have been 
pointed out, and only regret that so highly deserving a scholar does not deter- 
mine to forsake a way which, I am firmly convinced, is an erroneous one. 

On Phonology in Irish. 137 

sage. The first form contains consequently, one preposition fo-, 
after which the media was aspirated, the other several preposi- 
tions fo- ad- and do- fo- od, whilst d dropped, after it had 
changed the following media into tenuis, which however again 

DO , O 

sunk to a media in Middle Irish ; the gh in fagh is, according to 
this, mere root anlaut, the g in fag and tog,''u\ which the fusion 
of several prepositions is also indicated by the length, is the 
softening of the c which has arisen from dg. The form gheibhim 
(I find), given as a parallel form to faghaim, shows by the 
aspirated anlaut, which clearly distinguishes it from gabhahn 
(I take), O. Ir. gabimm-se (accipio, sumo), the loss of a preposi- 
tion ending with a vowel, perhaps fo-, for do- in an-dorogbid (gl. 
donantes), in Z. 1042, produces a different meaning; on the other 
hand, in bheirim (I give), parallel form of tabhraim, likewise 
plainly distinguished from beirim (I bear), by the anlaut, do- 
appears to have fallen off, for already O. Ir. dobiur along with 
tabur, i. e., do- fo- bur (do) exists. In deirim (I say), also, just 
as in the above-mentioned forms, a bh has been dropped, which 
is still retained in the perfect dubhras (Keating — dubhari), and 
is confirmed by the O. Ir. do-m-ber-som (quae dicit ille) ; the 
imperative abair (Mid. Ir. still apair O'D. 239), and the so- 
called conjuctive go-n-abraim, on the other hand, contain the 
same root ber lm combined with another preposition (aith- Z. 80) 
cf. epiur epur (dico), apir (dicis) atbeir adbeir epeir epir (dicit), 
also dianaiper (de quo dicit), Z. 1068, dian-eprem (de quo dici- 
mus), and many other forms in Z. to which nadipru, nadipro 
(who would not speak), Tir. in O'D. 436, instead of nad-idbru, 
appear also to connect themselves. Less striking disfigurations, 
but still sufficiently great to warn us of the necessity of extreme 
caution and moderation in the use of Modern Irish, are, for 
example, the softening of tenues to mediae almost everywhere 
in the inlaut, but even in the anlaut in get (what), gibe (whoever), 
gach (each, every), gan (without), go (to, with), and go (that), 
with the part. verb, gur, instead o£ cia (quid?), cip e, cib e (qui- 
cunque), each (omnis), cen (sine), co (ad, cum), co (donee, ut) 
and coro; the loss of the initial/in ri (with), and ar (upon), for /W 
(7rpoc) and for (super) 107 , which is probably only a continuation 
and repetition of an older phonetic process, so that a change into 

106 Cf. Skr. bru, Zend mru, Gr. pep and Fpe (spew, prjrcjp), Lat. ver-bum, Goth. 

107 In the Modern Irish ar, the two prepositions ar and for are so mixed up that 
it is difficult in each particular case to determine which of them we have to deal 
with ; the forms with suffixed pronouns undoubtedly contain for, and not ar : orm, 
ort, air, uirre or uirri, orrainn, orraibh, orra, or ortha, as evidently results from a 
comparison of the Old Irish— -form (more correctly formrri), fort, foir, fair, fuiri, 
furnn (Z. 1005) fornn form, foirib fuirib furib, form (c. d.foraibforib), on the 

138 EbeVs Celtic Studies. 

/immediately preceded the frequent loss of the p, thus for ex- 
ample patar may have first changed into fatar, and then into 
athir; loss of a vowel in da ("of which", also "which" and "if") 
for dian (from do-an, cf. Z. 892), the auslaut of which may still 
be recognized in the eclipse following ; consonantal metathesis in 
bearla, beurla for belre (lingua, sermo), for which berli is once 
found in Z. 9, in baistim for baitsimm (baptizo), eistim for eitsimm 
(ausculto), easbog (Mid. Irish easpog) for epscop, Cornish, escop 

However necessary in such cases we may find the Old Irish 
in the elucidation of the Modern Irish forms, and however 
clearly we may thereby discern the error into which the direct 
comparison of the latter with those of the other languages might 
lead us, the comparison of the newer forms is not less instructive 
and important for correctly understanding the older ones, nay, is 
often indispensably necessary, and a closer attention to those 
forms would have saved Zeuss from many errors. As sufficient 
preliminary investigations have not yet been made to render it 
possible to give a systematic representation of Irish phonology, 
I shall only touch in the following pages upon a few points to 
which my studies have led me. 

§. 2. Vocalismus. 

The most difficult part of the Irish phonetic system to bring 
to a fixed standard, is the Irish vocalismus, because three kinds of 
e and o appear to exist, which do not always admit of being dis- 
tinguished with certainty, and further, because even the question 
of the priority of a or o, a or e, u or o, i or e in individual cases 
is oiten beset with insuperable difficulties (at least for the pre- 
sent). In order to indicate graphically the threefold genesis of 
the e and o without the use of new type, I propose, firstly, to 
leave the e and o, which have arisen directly from a without the 
action of another vowel, unmarked, equally whether they 
sounded e and o in Gaulish, or came into existence later by the 
simple weakening of a (perhaps in the auslaut from 6 and 6?); 
secondly, to mark the umlaut caused by i and u with the sign of 
shortness, by which we gain at once a sign for original and 
secondary i and u, for ai and au m diphthongal and such as arises 
from umlaut ; and lastly, to denote the breaking by a, especially 

one hand, and dirium, erut-su, airi (the feminine does not occur), erunn, dirib 
airiuib-si, airriu erriu erru on the other. For the only deviating form O'Dono- 
van adduces Middle Irish, for raind, with which orrainn accurately agrees. 

108 Perhaps the most convenient way would he just to write this umlaut every- 
where ai, au. This mode of marking appears to me to be very convenient for 
Zend also, in order to distinguish the i and u in gairi, tauruna, from the original 
in gdus. 

On Phonology in Irish. 139 

weakenings from i and u by e and 6, the former to be under- 
stood completely in the sense of the M. H. G. e, the latter, how- 
ever, in the opposite sense of 6 in O. Norse. 

Examples: 1. O. Celtic e in ech (equus), Gaul, epo-, W. 3. 
ebawl, V. ebol (pullus) ; breth (judicium), Gaul, vergo-bretus; nert 
(virtus), Gaul. Nerto-marus, Esu-nertus, W. C. nerth, Arm. nerz; 
nemed i. e. nemedh (sacellum), Gaul, vefinrov i. e. vb/ustov, Ver- 
nemetis, W. 2. neuat, 3. neuad (aula) ; O. Celtic o in orcaid (occi- 
dit), orcas (qui occidit), i. e. org-, Gaul. Orgeto-rix, W. 1. orgiat 
(caesor) ; ocJit (octo), Gaul. Octo-darus, W. 2. uith, 3. wyth; — 2, 
umlaut by i — aith-, aid-, ed-, ith-, id-, Gaul, ate-, Kymr. at-, et- 
(perhaps also W. 2 ed-, e-, 3. yd-, y-, P. y-, Arm. ez-, e-, the 
verbal particle, = Skr. ati ?) ; air-, er-, ir-, Gaul, are-, Kymr 
ar-; erbaid (committit), erbid (tradite), root arb; umlaut by u — 
rolaumur, rolomor (audeo) — more frequently ai (oi) and au — 
baill, boill, baull, baullu, billlu; 3. breaking of i — etha, betha, etal, 
cenn, tuisiil; breaking of u — mo'ga, loth (lutum), crochad; simple 
weakening perhaps in felsitb, cruch, domun (Gaul, dumno-)? In 
order to distinguish ia = e and na = 6 from contracted ia and ua, 
I mark the former with the grave accent on a — e. g., Mad, (victus, 
esca) from *bivatha (fiiorog), hence gen. biith Mid, Mod. Ir. Mali, 
on the other hand, dia (deus), from *deva, gen. dei, de, so also 
uathath, uathad (singularis) ■=. othad. 

Even though it be established that the a in Mid. Irish mara 
(maris), mainistrech (monasterii) is corrupted from the o of 
O. Irish mora, monistre (monasteriorum), and the same observa- 
tion very probably applies to the Mid. and Mod. Ir. a of many 
endings (e.g., part. pass, in -ta, O. Ir. te) as compared with O. 
Ir. e, we cannot thence by any means conclude that this is every- 
where the case ; thus, for instance, that in the gen. sing, betho 
(mundi), we have an older form than betha m , in aecaillse, an older 
form than aecolsa (ecclesiae). The Mod. Ir. affords us little help 
in this investigation, because the uncertainty of the O. Ir. ortho- 
graphy (which, for example, leaves the umlaut of the a by i at 
one time unmarked, and at other times writes it ai, oi, ui, e, i, 
and even ae and oi) is not only in great part retained here 
(leaving out of consideration the action of the well known rule 
— caol le caol, leathan le leathan — a rule which, however, in its 
turn acts disturbingly), but also by arbitrarily confounding the 
simple vowels, has reached so great an extension that almost any 
short vowel may stand for every other. Thus a is found for u 
in chugam for cuccumm (ad me) in the ace. pi. chicca for cuccu (ad 
eos) as in Middle Irish already ; a for i especially before n (ana- 

109 Although Mr. Stokes, in his valuable Irish Glosses, p. 159, appeals for it to 
the ogamic gen. Atilogdo or Apilogdo. 

140 EbeTs Celtic Studies. 

logous to tlie French pronunciation of £?i=:Lat. m, in dans = 
de intus, sangHer=smgvlari.s even written), in the article an = 
ind, ant = int uo , in the preposition a(n)=ih, in the interrogative 
particle an(n)=in, while the prefix in- or ion- has preserved the 
i of the old ind- in colann (a body) =colinn (caro); even u for 
i in the preposition um=imm (also with suffixes iimam, etc.); 
o for 6 in romam, etc., roimpe^remi. In spite of this confusion 
in the elements, which for the e}^e is considerably increased by 
the well known rule according to which fear is written for fer 
(gramen), fear for fer (vir), fedrr for ferr (melior), even neoch, 
noch for nech (qui, properly aliquis), the Modern Irish comes to 
our aid even in the vocahsmus, whenever we have to do with 
the explanation and origin of true or apparent diphthongs. The 
O. Ir. ai has, for example, a threefold meaning, as a true diph- 
thong, as umlaut from d, and as umlaut from a; the usual mode 
of marking these in MSS. is not sufficient to properly separate 
these three sounds according to their different origin, the diph- 
thong appearing at one time with, and at another without, an 
accent, being consequently not sufficiently distinguished from 
either the short or the long umlaut. (In the marking of the 
umlaut by ai and di above proposed, the accent for the diphthong 
ai may be dispensed with). The parallel forms also (pi, ae, oe, 
for the diphthong, e, i for the umlaut at, a without umlaut for d%) 
do not give full security, for ae is sometimes found for e, namely 
in anlauts and auslauts, and di and oi sometimes for the umlaut 
ai, especially before liquids (Zeuss, 32). But if we compare Mod. 
Irish, the diphthong ai, ae appeal's transformed into ao (or its 
umlaut aoi) : caora =. caira (ovis), caoin = cdin (bonus), gaoth 
(already Mid. Ir.)= gdiih gdid (ventus), maoin (wealth) = 0. Ir. 
pi. maini (opes, pretiosa, dona), saobh (bad, evil) = s«i&, sdeb, 
soib soeb (falsus), saoghal (world) =saigul, maor (steward) = Mid. 
Ir. timer; the umlauts, on the other hand, have remained un- 
changed, maith (bonus), ainm=ainmm (nomen), aimsear = aimser 
(tempus), cailleach (a hag):=caillech (anus, monacha), gabhdil 
=gabdil (sumptio), except that, as already hi O. Irish, oi fre- 
quently occurs for a%, and seldomer ei — coill (wood) = caill (sil- 
va), clomne = clainne (prolis), anoir = anair (easterly), eile oile = 
aile (alius). The Modern Irish does not suffice, however, to 
distinguish ai and oi, for it expresses both by ao (aoi) e. g., aon 
(unus)=o#w oin, caol=zcoil (macer), coaga = coica (quinquaginta) 
— cuig = coTc (quinque), is remarkable. The Kymric dialects 
which have retained the ai, ae, as for instance the Welsh, but 
change oi, oe into u (with few, perhaps, apparent exceptions), 

■ 10 The explanation of the newer form -which I have attempted at p. 88, is 
incorrect, hecanse this phonetic peculiarity of the Modern Irish had escaped me. 

On Phonology in Irish. 141 

e.g., un — Irish oen, may be here appealed to. Umlaut ail and 
diphthong au (au, ou, do, 6) appear to be less sharply distin- 
guished, as the former is replaced by u or o, and the latter by 
6 or u, which is sometimes shortened, or its length is not 
marked, cf. aue (nepos) and o ua, augtortds (auctoritas), and 
ughdar, pronounced udar (auctor) ; the inorganic ail instead of 
ai in aud-, aur* (Z. 7. 8) does not occur at all in Modern Irish. 

§. 3. Consonantismus — Aspiration of Media? after Vowels. 

The comparison of the newer forms yields us much more impor- 
tant service in the consonants. Thus, for instance, at p. 119, a 
formse (already proposed by Stokes, Beitrage I. 450) for the ace. 
fern, siu for the ace. pi. was deduced from inte (in earn), intiu 
(in eos), airriu (propter eos) ; etarru (inter eos), form (super 
eos), and the Modern Irish which has only preserved the dative 
after di and do (diobh, doibh), but otherwise puts the accusative 
everywhere, offers proofs in abundance which confirm this con- 
clusion. The s of se and siu, su is preserved in thdirse, thdirsi 
(over her) and thdrsa (over them) ; it has changed into t after s 
in the secondary form thdrsta and in aiste, -ti (out of her), asta 
(out of them), likewise in uaiste, -ti (above her), uasta (above 
them), in which consequently s oi. st is to be considered as origi- 
nally ss (the original auslaut of tar(s) is perhaps still to be recog- 
nized in the rr of thorrainn, thorraibh, -orrainn, orraibh from 
for occur also, however, and the O. Ir. torunn has single r) ; th 
for s after vocalic auslaut in fuithe, -thi (under her) and fiitha 
(under them), uaithe, -thi (from her), and uatha (from them), 
trithe, thi (through her) and triotha ^through them), similarly 
after r in the secondary form ortha; rr for rs in idrre, -ri (on 
her), orra (on them), eatorra (between them) where at the same 
time the depressed tenuis in eidir is preserved; original tenuis 
preserved by s in aice, -i (with her) and aca (with them), chuice, 
-i (unto her), and chuca (imto them), while aige (with him) and 
chuige (unto him), prove vocal anlaut by the media; tenuis after 
nasals derived from an original media in uimpe, -i (about her), 
umpa about them, from a secondary one (?) in innte, -i (in her), 
ionnta (in them), directly intercalated in roimpe, -i (before her) 
and rompa (before them), on the other hand roime (before him). 
All these examples are in the highest degree important and 
interesting by the constancy with which the s, which has else- 
where generally disappeared, makes its influence still left in the 
latest language period, and most strikingly of all in roimpe (for 
roimsi the p as in Lat. dempsi, demptum) along with roime with 
aspirated m. 

But we especially want very often the Modern Irish to deter- 

142 EbeTs Celtic Studies. 

mine whether tenuis or media is to be read aspirated or not. 
As is well known, the oldest documents do not always very accu- 
rately mark the aspiration even in tenues, still less in the case of 
f and s, and not at all as a rule in the case of ruedias and m, or 
at most mark the unaspirated pronunciation by duplication, and 
in the MSS. of Zeuss, wherever the aspiration is indicated, the 
aspirated tenuis is found for the media. Thence arises a double 
ambiguity, inasmuch as we may fluctuate equally between d and 
dli as between dh and tli; but this ambiguity is still further in 
creased by the circumstance that tenuis not only occurs for double 
media, but also inversely media here and there for pure tenuis. 
As the Middle Irish MSS. also do not always accurately mark 
the aspiration of the media, it is often only the Modem Irish 
which can here help us, for the latter, in spite of the above men- 
tioned corruption, has, by completely dropping aspirated conso- 
nants, and a wide spread lowering of pure as well as aspirated 
tenues, fortunately maintained accurately, on the whole, the 
limits between aspiration and pure pronunciation, with the excep- 
tion of some verbal forms before which particles have dropped, 
and some particles whose anlaut is aspirated as, cheana (already), 
bheos, fds (yet) for cene (jam), heos (adhuc), co and tar in the 
formulae cliugam, thorm (cf above.) 

The simple m which in O. Irish is not protected by consonants, 
becomes always aspirated in Modern Irish ; ^domain mfudumain, 
fudomain (profundus) becomes doimin (although the second 
vowel was probably intercalated here merely to ease the pronun- 
ciation on the dropping of the O. Irish ending), and this m has 
likewise (even in Middle Irish) frequently taken the place of an 
original bh as in naorh instead of noib (sanctus), neam-, neim- in- 
stead of neb- (negative prefix), claidhem instead of claideb (gla- 
dius), fealsam instead of felsub, which has in consequence fol- 
lowed the false analogy of brithem. We may, therefore, with 
perfect security deduce from m in inlaut in Modem Irish, m or 
mm (mb) in Old Irish, which to be sure we are not as yet always 
able to explain; thus anam points back to animm (anima) Z. 
1059, ainm to ainm (nomen), uaim to uaimm (a me), etc., as im, 
trim, urn, does to imm (imb) ni while dom has become dam (to 
me), rem roim (before). 

An original media after vowels is always aspirated in Modern 

111 As in imm from mb, so may the m=in in Cormac (for ogamic Corpimaguas, 
where the vowel dropped should produce aspiration) have been assimilated from 
pm, in ammi (sumus) from sm; in animm, ainm, uaimm it is just as unsatisfac- 
torily explained as in 1 sg. and pi. of the verb : and singularly enough the Kymric 
shows just here a softening, V. enef Arm. enef, eneff] (anima), W. 3. enw, P. 
(Ji)anoic, Arm. hanu (nomen), "W. 3. ohonqf, ahanaff(a. me, de me), just as in 

On Phonology in Irish. 143 

Irish, adharc, brdgha, buidhe, croidhe, a n-deaghaidh, foghlaim, 
adhradh, gen. adhartha; therefore, no doubt, to be thus re- 
presented in Old Irish: adarc (cornu), brdge (cervix), blade 
(flavus), cride (cor), indegaid-h (post), foglaim (cornprehensio), 
adrad (adoratio). The change between aspirated tenues and 
mediae also points in the same direction; the final med. asp. is a 
softening from ten. asp. in adrad (probably also in indegaid) as fre- 
quently happens, cf. cailleach, Old Ir. cailltch, gen. caillighe. After 
consonants the mediae in Modern Irish also remain without aspi- 
ration, except where a vowel has dropped out, drd, fearg, bolg, 
borb, O. Ir. — ardd (altus),/erc for fergg (ira), bolc^bolgg (bulga), 
borp, i.e. borb (stultus); Stokes (Beitrage II. 102) has, therefore, 
rightly looked upon such forms as dealbh, marbh, tarbh, where 
the mediae appear aspirated after liquids, as proving bh = v. n2 

On the other hand, the mediae are often assimilated after 
liquids, especially after m and n [as partly already in O. Irish, 
uall (superbia), gen. uailbe], thus in agallam = acaldam, accal- 
da?n, acaltam, i. e., accalddam (allocutio), iomad Corm. Glos. 
(many) z= imbed (copia, ops), ionamz^indiumm (in me), binn 
(melodious) = bind, clannz=. eland (proles), cunradh, Mid. Ir. cun- 
dradh (a covenant), O. Ir. eundrad (merx), connavcas (I saw), 
for cond. (root dare in SlpKw, etc.), coinneal (a candle), cf. eain- 
dloir (candelarius), even Middle Irish bennacht, bennachadhz=. 
bendacht, bendachad (benedictio) likewise mallacht-=maldacht 

The so-called eclipse also depends upon the assimilation, so 
far as it affects mediae, inasmuch as na-m-ball (membrorum) is 
pronounced nammall. I suspect, therefore, that in O. Irish also 
the dot over h and Wi before mediae had more to do with the 
media than witjr the nasal, and consequently that rad ridd is to 
be pronounced rad ne (notwithstanding the apparently contra- 
dictory mode of writing frecdairc, dofoirde), because nasals other- 
wise regularly drop out before tenues, but not before mediae, or 
rather remain when tenuis becomes media, as in ind- (Gaul. 
ande-), ingor (Lat. ancora). Another assimilation according to 
which codhladh (sleep), ceadna (the same), colna (of the flesh), 
are pronounced colladh, cSana, colla, is not indicated in writing. 

§. 4. Consonantismus — Aspiration of Tenues after Vowels. 
The original tenues (and the hard spirants s, f) like the 
mediae, are always aspirated in true Celtic words after vowels, if 

1 sg. -of, while although an Arm. dif, diff corresponds to the Irish dom, dam, we 
have on the other hand W. 3. im, ym, P. thym. From this it appears that the mm 
in these cases is exactly comparable with the nn of the article, and was perhaps 
produced under the influence of the original accent. 

112 Derbh (certus) along with dearbh is very curious, so likewise is easbha 
(defect), pi. gen. easbhadh, cf. ace. tesbaid (defectum), dat. tesbaith. 

144 EbeTs Celtic Studies. 

a vowel or liquid follows, but not before mutes, except in the 
combination ct, which is sometimes written cht, also, as it ap- 
pears, not before (dropped) v; and in this the Modern Irish has 
altered nothing, except that it has logically carried out the cht; 
with oscillation in Old Irish of aspirated tenues to mediae, especi- 
ally in th, less so in ch, in which latter in Modern Irish it has 
much more extensively spread. After consonants (as before 
mutes) tenuis remains without aspiration, also after those which 
have dropped out, hence t, c, instead of nt, nc (likewise /, s, 
instead of nf, ns); but Modern Irish has here frequently lowered 
the tenuis to media, both original and secondary. 

The old Irish has changed organic mediae into tenues in two 
ways: 1. before dropped vowels, by which the media has to a 
certain extent passed into auslaut, and thus become hardened to 
tenuis, for example in tdirci (efficit) from do-dtrci, in the com- 
pound prefixes int- from in-do-, tair- from do-air- (flair-) taith- 
from do-aith-, tes- from do-es-, tiar- from do-iar-, timm- from 
do-imm-, tin- from do-in-, tind- from do-ind-, the same with the 

dropping of an /in tu-, to- from do-fu-, do-fo-, in tor-, tor- (tuar-, 
tur-, ter-) from do-for, with the dropping of an s in intsamail, 
intsliiiclit, in the article int- from ind-'s-, and in the abovemen- 
tioned prepositions with suffixed pronouns ; 2. by the collision 
of two mutes, in which the first, if it was a media, became 
on that account hardened, and then induced the hardening of the 
second, just as if it was an original tenuis or aspirate, atomaig 
from ad-dom-aig , cotondelcfam from cot-don-delcfam (cot- accord- 
ing to Stokes, Beitrage II. 106 — Welsh cant-), fritammiurat 
from frith-damm-iurat, and others given by Zeuss, 336, edpart, 
idpart from aith-bart, but has then generally been dropped, 
or more correctly has assimilated itself (for gemination often 
remains unexpressed in O. Irish, and in the case of consonants 
capable of aspiration, always in Modern Irish, only 11, nn, rr 
are written), thus in acaldam accaldam (allocutio) from acl-gal- 
dam u3 (pronounced atgaldam, atcaldam), epil (pent) along with 
aibail Z. 1012 (pronounced atpail) from aith-bail, ecne (cognita) 
along with aidgne aith-gne, frecre (responsum) from fritli-g(a)re, 
conucbad (ut attolleret) from conuad-gabad, doopir (privat, aufert) 
from do-od-bir. 

In the first case the Modern Irish preserves the tenuis which 
is thus produced, e.g. in tim-, in the article ant and in the above 
examples of prepositions with pronouns ; in the second it allows 

113 Cf. adglddur (jrpoaayoQivia), adglddathar (appellatur) ; so also comalnad 
(impletio) along with Ian (plenus). The abovementioned hardening is also, no 
doubt, the reason for the mode of writiDg^, dd, bb, for c, t, p. 

On Phonology in Irish. 145 

the same tenuis (the second mute) to again sink to a media, but 
does not aspirate it, e.g., iodhbairt (an offering) = idpairt, agallain 
(a dialogue) = acaldam, ei-blim (I die), likewise *eptimm, eagna 
(wisdoni):=£cfte (sapientia), admidm (I confess) cf. ataimet (profi- 
tentur) from ad-daimet; both united show themselves in the 

abovementioned togbhaim, where the t of tocbaimm from do-fo- 
od-g) has remained, but the c has sunk to g. It has likewise 
changed the original tenues, to which (n)t and t(v) consequently 
belong, everywhere into mediae after vowels: codladh (sleep) = 
cotlad (somnus) dat. cotlud, Z., S22, fad=fot (longitudo), cead=: 
cet (centum), ceadna (the same) = cetne (primus), creidim = cretim 
(credo), sometimes even geminated ones as, for instance, clog= 
clocc (clocca), beag (little) = becc bee (parvus, paucum), along with 
these there are however mac = mace (films), cnoc (a hill) — cnocc 
(gibber, ulcus), also emit (a harp) =crot, i. e., crott (crotta), breac 
(a trout), gen. brie, which points to *brecc (cf N.H.G. briche); 
trocaire (misercordia) from trog-caire (amor miseri) also remains 
unchanged. Fluctuations occur here after consonants; after s 
generally softening ; less frequently and more properly in Gaelic 
after eh (after gh, — ughdar, O. Irish augtortds); after I and r the 
tenuis is preserved — ole, marc, mart, fait, corp; but p often 
passes into b after l Ui \_Alpa, gen. Alpan Cormac's Gloss, in 
O'Don. 3. 354 (Scotland), ace. Alpai-n (Alpes) Z. 616, from 
which cenalpande (cisalpinus), therefore properly "highland", has 
become Alba, already Middle Irish gen. Alban in O'Don. 83, dat. 
Albain 251], less frequently after r (yet carbad = carpai), t re- 
mains also after n in muintir, muintear, but c passes into g — 
rangas (I reached), thdngas (I came), in Middle Irish still rdnca- 
tar (they reached), O'Don. 246, tdncamar (we have come), 252. 
It is evident that the so-called eclipse of the tenuis, and 
of /, which sinks to bli under similar conditions, (strictly speak- 
ing no eclipse can be spoken of in the case of s, as the t before it 
belongs to the article, otherwise we would be obliged to consider 
the p of umpa to belong to the eclipse) also depends upon this 
sinking to mecliag, and has properly nothing whatever to do with 
the nasal, which is generally dropped before it. Just as in the 
middle the tenuis has changed into a media indifferently, whether 
a nasal has fallen out before it or not, as cet, etar become cead, 
eidir, exactly as bee becomes beag, the former is, however, acci- 
dentally the more frequent, so in the anlaut, imder certain con- 
ditions, every tenuis not protected by consonants also passes into 
a media, and it is a simple accident that in most cases a nasal 
originally preceded, and that consequently, as a rule, the funda- 

114 Probably the bh in dearbh, easbha maybe thence explained ; see note 112, 
p. 142. 

146 EbeVs Celtic Studies. 


mentally different eclipses of the tenues and mediae go hand in 
hand; that this is not a necessary condition is shown by the 
eclipse after ea-, ei- (O. Ir. e- along with es-, like Lat. e along 
with ex), which only occurs with tenues, eagcdir (injustice) = 
ecoir (incongruus) , eadtrom Qight) = Strum (levis), not with me- 
diae, eadoimin (shallow) , because no nasal is present. 

From what has been said above, we may consequently con- 
clude with perfect safety that Modern Irish tenuis corresponds to 
O. Irish tenuis, Modern Irish dura to O. Irish dura, on the other 
hand aspirates to aspirates with exceptions, Modern Irish mediae 
to Old Irish mediae only if aspirated, or in the combinations rd, 
lg, rg {Id and nd have been assimilated to 11, nn), while after 
vowels, s and ch every pure media points to an old tenuis, after 
I and r at least b is of uncertain origin. We may therefore in- 
fer from drd ardd (sublimis) — written ardd, art, ard, from fearg 
fergg — written fere, from bolg bolgg — also written bole, likewise 
from agallam accalddam — written acaldam and acaltam, from 
binn bindd; on the other hand borb would not lead with cer- 
tainty to borbb, if we did not find burbe written along with 
burpe. The circumstance that dura point back to dura will, 
however, be of especial use to us in the case of dentals, for the 
purpose of getting rid of some errors into which Zeuss has fallen 
in several passages of his grammar, in consequence of having ne- 
glected the newer forms. 

§. 5. Consonantismus — Cases ivhich afford occasion for Aspira- 
tion after a preserved or lost Vowel: (I.) in Inlaid; (II.) in 
Anlaut; (III.) in Syntax. 

As is well known, the same laws which govern aspirations 
after vowels, apply in general to those cases also where vowels 
had originally existed, but dropped out, so that we may infer 
from the appearance of aspiration the former presence of a vowel 
in inlaut as in anlaut ; if, therefore, for example, s before mutes, 
(according to O'Don., also before m, cf. fosmachtu, Z. 6Q6, con- 
sequently before consonants capable of aspiration generally) be 
not infected by preceding vowels, as the mode of writing tesst 
shows, a doinscann-som, intinnscana (incipit) from do-incF-sc., in- 
do-ind'-sc, will stand opposed perfectly according to rule to the r 
inthamuil, intsliucJit from ind's. The aspiration rule is, how- 
ever, subject to so many exceptions in this case, inasmuch as it 
also depends upon the nature of the preceding consonants, that 
in the uncertainty of the ancient orthography we can only attain 
safe results by a comparison of all individual cases with constant 
reference to Modern Irish. 

Such cases as afford occasion for aspiration by a preserved 
or lost vowel, belong essentially to three categories : 

On Phonology in Irish. 147 

(I.) In Inlaut. In the inlaut of a word before, and in the 
derivative or flexional endings, especially in the word-forming 
suffixes -ath, -eth, -uth, -id, -ach, -ech, -ithe (-Tde), and before the 
-t (th, d) of different conjugational endings. In all these cases 
occasion also often occurs for the dropping of a vowel in 
inflexion and derivation, and Zeuss (page 84, with which the 
examples 762 seq. may be compared) has correctly remarked 
that "the t of the ending is not aspirated after I, n, s, and that 
a tt (or t) arises from t-t, th-f. 

The following examples are from the conjugation: con-festa 
(ut scias), marufeste marrnfeste (si sciretis), condigente (faceretis), 
nigette (Z. 264, "non faceretis"?) conrochretesi (concrederetis), 
conndruchretesi (ne crederetis) — with t for tt — along with nis- 
cartha (non abesses), nongabthe (q. sumebatis), fut. secund. fol- 
nibthe Z. 454; deponentials — rofestar (scit) nifiastar (nescit), 
miastar (judicat) and the preterites in -astar, -istir, -estar; pas- 
sive forms — arna fur astar (ne fuscetur), samaltir (comparatur), 
adcomaltar (conjungitur), clonelltar (q. declinatur), manireltar 
(nisi manifestatur), frisduntar (obstruitur), asagnintar (signifi- 
catur), gentar, do-gentar (fit, fiet), nomglantar (,,emungor") non- 
lintarni (iruplenrur) , nonnertarni (q. confortamur) for tt, con- 
intorgditar (ut non circumveniamur) and honuntogaitami (ex 
quo fraudamur), sluintir (signiflcatur) with t for dt, on the other 
hand derbthair, scribthar oinaichthir, cairigthir, lobrigthir, sui- 
digthir, intoichther, indtuigther, arosailcther, a-carthar, itarscar- 
thar, anasberthar, asrirther, fristacairther, berthir (differently 
nomthachtar („angor") and genthir, Z. 470 !) ; preterites — 
doronta (facta sunt), asridenta (inquinata sunt) along with dorur- 
gabtha (prolata sunt) ; secondary tenses — nolintae (solebat 
repleri), conulintae (ut compleretur) along with arna eperthe, 
doberrthe, roberrthe, nocrochthe, na ructhae; past participles — 
accomallte acomoltae (conjunctus), comchlante (conseminatus) 
with t for dt, remfoiti (praemissi) so also dlutai ace. pi. (flxa) 
1015 for tlit., forbanda (secta) 845 with d after n. On the other 
hand, remeperthe, sulbairichthe, atdchutmthe, loiscthe, aurgabtlia, 
(timmorte is curious with the c dropped as in the preterite 
dobimchornartt, further imdibthe (circumcisus)and/o« , c^/ie (erudi- 
tus) 115 where, after the loss of the n of ben- and can-, we should 
expect uninfected t, forngarti (jussi) appears like timmorte to be 
formed without a copulative vowel) ; future participle — eclustai, 
sastai, imcasti, airillti, denti, forcanti, cocarti, for cocartti (einen- 
dandum), in opposition to eperthi, imcabthi, (aichti is curious !). 
The whole of the examples, with the exception of the evidently 

115 Imdibthe and foircthe may be compared with Sanskrit and Greek forms, 
such as hata, (parog, from han, (pev. 

148 Ebel's Celtic Studies. 

corrupt genthir, confirms throughout the observation of Zeuss ; the 
omission of the aspiration takes place only after I, n, s, d, t, th, in 
opposition with crochthe, among others, except in the case of nom- 
thacthar and aichti (timmorte and forngarti may be explained in 
this way, that these verbs go in accordance with series III. of 
Zeuss) ; it is therefore singular that O'Donovan, in the rule for the 
Modern Irish passive and participles, puts tenuis after all aspirates 
ch, gh, th, dh (others do not here occur), except in the verbs in 
-ighim, as well as after I, 11, n, nn, s, while, on the other hand, he 
puts the aspirate after d and t. He at the same time admits, how- 
ever, that the sound remains the same after d and t, whether we 
write t or th. This rule also receives no confirmation otherwise, 
inasmuch as t is everywhere found in derivation and flexion both 
in Old and Modern Irish after I, n, s, t, d, th, dh (only with 
softening in d after n, seldomer after /), on the other hand th 
appears equally constant after ch, gh, as after all other mutes. 

The suffix -tu masc, -atu, -itu (cf. Beitr. II. 81), seldomer -ti, 
especially affords us examples from the declension, as it is usually 
affixed without a copulative (hence tabairt, epert): gen. pectha, 
pectho, nom. plur. pecthi, pecthe, pectha, gen. pecthe, dat. pecthib, 
&cc. jiecthu (pectha Z.1003) fromjieccad [i.e. peccatli) , gen. cro'chtho 
from crochad, e'tarscartha from e'tarscarad, cursagtha from ciirsa- 

gad, d&nigikea from daiugud, foilsichtho indfoilsTgthe from fotl- 
sigucl, incholnXchtlio inchohugthta from inchohugud, intsechtaigtha 
(read ints.) from sechtaigud, sulbarrichthe Z. 618 froni sulbair- 
igud; gen. iarfaichtheo iarfaigtho, dat. pi. icrrfaigthib Z. 1070 
from iarfafgid, iarjXgid f., dat. pi., debthib from debuith; on the 
other hand, gen. relto from relath, relad (manifestatio) , ind- 
aerchoiltt'a from erchmliud (deflnitio), cesta cesto from cesath 
chad (passio), nerta from nertad (exhortatio) , tairmchrutto from 
*tairmchruthad (transformatio) , gen. dag-imrdta, drog-imrdto 
(it is to be read thus), nom. pi. imbrdti imrdti, ace. imrdtiu, 
(Z. 1068), from imbrddud imrddud (cogitatio). 116 Here also t 
remains after I, n, s, and dental mutes, but is aspirated after all 
other consonants, and the Modern Irish confirms this by the 
plurals sgealta, sedlta, cedlta, nealta, bailte, coillte, aitheanta, 
leinte, teinnte, linnte, cluainte, mointe, tdinte, cointe, brointe, and 
the genitives ionganta, tionnsganta, cosanta, deanta from sgeal (a 
story), sedl (a sail), cedl (music), ileal (a cloud), baile (a town), 
coill (a wood), aiihne (a commandment), leine (a shin), teinne 
(fire), linn (a pool), cluain (a meadow), moin (a bog), tain 
(a flock), cu (a greyhound), bro (a quern), iongnadh (wonder), 

116 Zeuss, 851, erroneously assumes a nominative dagimrai. Stokes (Beitrage 
I. 450) also is in error respecting tairmchrutto (crochia appears to be careless 

On Phonology in Irish. 149 

tionnsgnadh (beginning), cosnadh (defence), dSanadh (doing), in 
opposition to the plurals murtha, cogtha, toirthe, teangtha, the 
genitives dadrtha, adhartha, cunnartha from mur (a wall), 
cogadh (war), toradh (fruit), teanga (a tongue), daoradh (con- 
demning), adhradh (adoration), cunnradh (a covenant), in which 
it makes no difference whether the suffix -at is originally word- 
forming as in teinne, or determinative as in cu. 

Derivatives in -te (i. e. -tia or -taja) after s, I, n, in which, how- 
ever, d appears generally after I and always after n (evidently 
pure d and not dh), see in Zeuss 763 seq.; whether, however, 
mistae (menstruus), conde (caninus), anmande (animalis), tal- 
mande (terrestris), eiscsende (,,intensivus"), cenalpande, aniendae, 
which are evidently derived from consonantal stems, have 
actually lost a vowel before the suffix, remains doubtful ; the d is 
to be read aspirated after r and other consonants as after vowels, 
bithgairddi (perpetuo breves), has been wrongly explained, like 
cethargarait, it belongs to an i- stem, and is to be further carried 
back to an nt- stem. To the examples for tt, t from ft, oVt, 
tlit, am-brotte (momentaneum) , gutte gutae (vocalis) — from 
which ahgutas 750 (vocalitatem suam) — , aicnete (naturalis), 
scote scotae (,,violarium") from scoth (a flower), are evidently 
to be added uathate (singularis) from uathath, from which ace. 
pi. fern, huathati, dat. pi. uathataib, and slabratae (the gloss 
catinensis being erroneous) from slabrad (catena), which Zeuss, 
769, erroneously places under -ant, so also, most probably, du- 
nattae (castrensis) from *diinad, cf. a righduinte (their royal 
forts), Cormac's Glossary in O'Don. 233, arsate (antiquarius), 
cf. arsid (a genitive as it appears) Zeuss, 581, plur. tuati (gen- 
tiles) 1043, from tucith (populus), perhaps also tecnate (domes- 
ticus) ; in the consonantel stems with the nom. -atu, -etu we may 
assume *-ntat, but they could also have arisen from *-tvat (cf. 
Skr. -tva n., Lith. -tuva m., Slav, -stvo n., but especially Lat. 
-tut f. in juventus, virtus, servitus, senectus), which is sufficient 
reason for their retaining the tenuis -t U7 as in the pronouns of the 
second person. Mod. Irish examples : saoghalta (worldly), gallda 
(exotic), fireanta (righteous), grianda (sunny), banda (feminine), 
also with assimilation daonna (human) = doinde; on the other 
hand, mordha (majestic), feardha (masculine), or dha (golden). 

Derivation with various suffixes : ecintech (inflnitus), from cin- 
niud (defmitio) huatigitir (rarescunt), from uathad, boltigetar 
(olent), from bolad, muntith (in^titutor), from munud; ingrentid 
(persecutor), lintidi (fartores), irchoiltith (maledicus),from irchol- 
lUd — esartaid (caesor) is remarkable, exactly like timmorte! — 

117 Nebmarbtu, -tath are at all events correct forms, and unjustly doubted by 
Zeuss, 763. 

150 EbeVs Celtic Studies. 

muntar (familia) is also, no doubt, to be placed under this category, 
and not to be compared with Gaulish ko/uovtoqioq ; centat (capi- 
tulum), from cenn, srdthataih, read -tat (aculeus), from srdthath; 
on the other hand, epertith, berrthaid, doilbthid, debthach, and 
dephthigim, tirthat, from tir, etc. — Compare the Modern Irish in- 
finitive, do chantain, but cFfearthain. — In diltuih, for example, 
the stem- vowel has been ejected, and because / precedes, we do 
not on this account know whether a mere vowel, or n, or a 
dental mute dropped with it. 

(II.) In Anlaut. In the anlaut of the second member of a 
compound, whether the first member be a noun, a numeral, 
or a particle, the second a noun or a verb. Neither here 
nor in the syntax has Zeuss brought together the exceptions 
to the aspiration rule ; but we may assume a priori, that the 
well ascertained law, according to which the dentals are not 
aspirated after I, n, s, t, d, th, dh, has in the main come into 
play also in composition and syntax, because it has a pure 
phonetic reason in the homorganeity of these consonants. 
Grimm (Geschichte der deutschen Sprache, 375) observes about 
Modern Irish " the Unguals t and d suffer, however, no aspira- 
tion after liquids, but remain unchanged"; but this is taken at 
once in too wide and too narrow a sense, for m and r do not 
hinder the aspiration, 118 and the mutes hinder it as well as n; 
what he further says, " I find also mactire, son of the land, as the 
poets call the wolf, not macthire", may be very simply ex- 
plained in this way, that this is not true composition, but merely 
juxta-position of the substantive with the governed genitive (= 
maqvas tirais), where there exists no reason for aspiration. Ac- 
cording to O'Donovan, 336 seq., aspiration does not occur (except 
in the case of s with a mute following, to which, according to p. 
54, we must also add m) with d, t after n, d, t; finally, in 
some cases not specially stated; I, s and the aspirated th, dh 
are not there mentioned, but it is scarcely to be doubted that 
they exert the same influence on d, t following, as we even find 
dall-ciach (a blinding fog) given without aspiration, so likewise 
athtkaoiseach (a deposed chieftain), and aithdheanam (remaking) ; 
but, however, aithteidhte (re-heated), athdoidhte (reburnt). Now, 
if even the Mod. Irish, in which aspiration is so widely spread, 
that it has come in after every particle in composition, with 
few exceptions 119 (ea- or ei-, eas-, con-, or coin-) has, nevertheless, 

118 Compare imdhiden (shelter, defence), urdhairc eardhairc (illustrious, re- 

119 The eclipse after di is perfectly enigmatical in diombuidheach (unthankful), 
diombuan (perishable), analogous to diomolaim (I dispraise) on the other hand, 
with aspiration, dfomor (very great), dicheannaim (I behead), diothoghluidhe 

On Phonology in Irish. 151 

preserved in the above position the d and I pure ; with much 
greater certainty may we look for the same thing in Old Irish, 
where the original limits of aspiration are exceeded only in very 
few instances (in du~ and mi- for *dus- and *mis-). Accord- 
ingly, we find 5 preserved before mutes in banscala (servae), cdin- 
scel (bonus nuntius), drogscela (malos nuntios), soscele (evange- 
lium), athscribend (rescriptum), incomscrib'ndaUh (syngraphum), 
dosceulaim (experior), doinscannsom (incipit), after the verbal 
participle in roscarsam (recessimus) ; t after n in — banter isrmd 
(obstetrix), grientairissem (solstituim), medontairismul (medias- 
tinus) — compare Mid. Ir. baintigerna (domina), in Stokes 1 Irish 
Glosses, — fintan (vinetum), cdinteist (bonum testimonium), cdin- 
toimtiu (bona cogitatio), cdintol (bona voluntas), caintaidlech 
(satisfactio), sentinni pi. (anus), intonnaigim (inundo), intursitib 
(irriguis), tintuth (interpretatio), fointreb (supellex) ; 120 after I in 
ind-idultaigae (fani), iltoimdden (cld = t, multarum opinionum), 
after s in rostdn (rosarium), after t in rechttcurcid (legislator), 
after th in frithtasgat (adversantur), frithtaidechtae (contradic- 
tionis), for which fritt*, frit- is also written ; we have, consequently, 
to consider d after n in banddlem (hospita), bandea (dea), ban- 
dechuin (diaconissae), bandachlach (leno) — cf. Mid. Ir. baindea in 
Stokes' Op. cit. — cdinduthracht (bona voluntas), sendmne (vetus 
homo), and after I in ilddni (multae artes), as dura, a hardening 
to t occurs after t, th, d (see supra), biddixnugnd, i.e., bithd., 
however occurs, Zeuss, 781. For some other exceptions, such as 
the above mentioned atbaU for athbail, idpart for idhbhart, where 
the hardening comes into play, at the same time (leaving out of 
consideration faulty spelling), I have not been able as yet to find 
any fixed rule ; only we must not take for an exception what is not 
one, as for instance the name Dunpeleder, Zeuss, 821, in which 
the p has remained pure, because this is no more a case of true 
composition than the above inactive, or the family names with O 
and Mac, which for the same reason are not. aspirated, e. g., 
O'Briain (gen. I Bhriain, dat. dTJa Bhriain, ace. ar O'Mbriain, 
according to O'Molloy, in O'Donovan, 369). 

(III.) In Syntax. In Syntax, the Modern Irish should be used 
only with the greatest caution for determining the laws of anlaut 
(which were not very clearly or completely developed by Zeuss), 
because it has here given way still more to the tendency to use 
this, originally a purely mechanical phonetic change, as a dyna- 

120 Zeuss indeed assumes (195,848), after in- also in composition unchanged an- 
laut, as, however, the n does not drop out anywhere, we must presuppose a 
fundamental form, like Greek hi, consequently aspiration which is supported by 
inchosc (significatio), etc. 


152 JEbeVs Celtic Studies 

mic agent, a tendency that was already visible in the particle 
composition, and arbitrary rules of scribes and grammarians, who, 
as a rule, had no idea of the nature of aspiration and eclipse, have 
had their share in still further disturbing and confusing the 
original rule ; nevertheless it may be here also of real service to 
us, if we consider perfectly unbiassed each grammatical form as 
that which it is, and not what it pretends to be, and bestow the 
necessary attention on the actual or apparent exceptions. 

The phonetic changes are dependent in Syntax on two con- 
ditions: not merely on the nature of the sounds which come 
together, but also on the greater or lesser logical correlation of 
the words, a condition which did not at all come into consideration 
as an independent one in inlaut and composition ; as in French, 
the pronunciation of the final consonant of the first word, even 
if it be capable of becoming sonant, does not take place between 
every two words, so in Irish, also, the auslaut does not exert in 
every position its influence upon the following anlaut. The 
closest combination is formed by the substantive with the pre- 
ceding article and pronominal genitive, the preposition with its 
case, the verb with particles and pronouns, which, in writing, 
are either enclitic or proclitic; the substantive is less closely con- 
nected with a succeeding adjective (as a rule, a preceding one 
enters into composition) still less with a dependent genitive, 
the connection of the verb with a substantive, as subject or 
object, is the loosest. Next to this, the very unequal action of 
the auslaut is of importance: original s- auslaut does not appa- 
rently lose its protecting action in any position in Old Irish, 
upon succeeding initial consonants; final n also occurs often, 
where the combination is by no means so particularly close, as 
in guidimse dia nerutsu; there are even phenomena which appear 
to point to an v ^{Xkvgtikov (cf. p. 90) ; on the other hand, 
the aspiration required by vocalic auslaut often does not occur 
even in the anlaut of the adjective, still more frequently in that 
of a dependent genitive. Many instances of omission are, of 
course, only a consequence of careless and imperfect writing, as, 
for instance, s and f often appear without a dot ; but others are 
due to perfectly determinate phonetic laws, especially to the two 
frequently mentioned above; this is especially seen after the 
article, by comparing the older and newer form. 

(1.) The article originally ended in s in the nom. sing, masc, 
gen. sing., and nom. plur. fern., dat. and ace. plur. of all genders 
(the neuter appears to have early passed over into feminine in 
the plural) ; in n in the nom. sing, neut., the ace. sing, and gen. 
plur. of all genders; in vowels in the gen. sing. masc. and 
neut., nom. plur. masc, dat. sing, and nom., and ace. dual of all 

On Phonology in Irish. 153 

genders ; according to this we have to expect in the ace. sing, 
(and nom. plur. nent.) and gen. plur. n (m) before vowels and 
medise, assimilation before liquids, and pure anlaut in the case 
of tenuis, s and/; in the gen. sing, masc, neut., nom. plur. masc. 
and dat. sing, aspiration, which is not written for medise, and in 
other instances pure anlaut of the following substantive or 
adjective. Most examples, also, agree with these observations, 
leaving out of consideration neglected aspiration, especially of s 
and/, which, however, in the case of s is generally made observ- 
able by a preceding t for d; gen. inspirto, intesa, etc., should not, 
however, be reckoned among negligences of writing, they are to 
be looked upon rather as actual exceptions, according to deter- 
minate laws. The s, in inspirto, cannot be aspirated on account of 
the following mute, hence the article is not written here either int- 
or ind-, because the t is hardened out of d or intercalated only 
before s (for which of the two explanations is the correct one, re- 
mains for the present still doubtful, as even in Modern Irish, roimpe, 
from roim si also appears along with uimpe, from utinU si) ; on this 
account insenduine does not get a t in the nom., but it does in the 
gen. intsenduini, inaccurately written indsenduini, and the pro- 
nouns sa (so, se) and sin, which are not aspirated, no doubt, be- 
cause a double consonant originally existed in anlaut, form every- 
where with the article inso insin (Zeuss, 275, 353, seq.). That 
the dental has also been preserved pure in intesa by the preced- 
ing sound (n or d), consequently that an aspirate is not to be pro- 
nounced here any more than in induini, and that Zeuss, 231, 
232, 236, with all the observations appertaining thereto, is 
decidedly in error, is shown, besides, by the constancy in the 
examples (nom. fern, indtogas, gen. masc. neut. intairmchrutio, 
intesa, intaidlich, dat. ontechtairiu, dontorud, isintnisiulsin, 
ontrediu, dindtrediu, iarsintairgiriu, hisiniorunt, hontecnataiu, 
dintecnatatu, issintodochidiu, isintuaichli, nom. masc. pi. intu- 
isil), also by the Modern Irish, which has maintained the same 

The laws of anlaut after the article look, to be sure, on first 
sight, and as represented by the grammarians, wonderful enough ; 
that the nom. fern, and gen. masc. cause aspiration, and the gen. 
pi. eclipse agrees with the old rule ; but that eclipse should 
occur in the dat. sing, after all prepositions, except do in West 
Munster, as G'Ponovan gives at p. 63, — and except do and de, as 
is stated at p. 393, — that s suffers the so-called eclipse onlv after 
do, de, is with the article, consequently aspiration with an inter- 
calated t before it, p. 70, — that t and d, as a rule, suffer in the sin- 
gular as little eclipse as aspiration, while in the genitive plural, 
on the other hand, they are regularly eclipsed, — appears enigma- 

12 B 

154 EbeVs Celtic Studies. 

tical, and the contradictory rules of the grammarians respecting 
the anlaut of adjectives (p. 110-117), appear to make the matter 
completely inexplicable. But if we examine this phenomenon 
closer, and compare the use of the prepositions and the examples 
from Keating (p. 394, seq.), light will be thrown upon this pecu- 
liarity, in which the confusion of speech among the people, and 
the foolish caprice of grammarians, have gone hand in hand, and 
immediately the exceptions become satisfactorily explicable. At 
p. 78 seq. (ante) attention has been before directed to the confu- 
sion in the case-endings, which had partially begun already in Old 
Irish, and which has been carried to an extreme in the Middle and 
Modern Irish ; we can now complete and correct what has been 
there said. In the first place, almost every distinction between 
nom. and ace. has disappeared, in the singular, the nominative 
form, in the plural at one time the latter, at another time the 
former, has alone been preserved, and even where in an isolated 
instance both forms occur, they appear to be promiscuously used ; 
the accusative form has very early replaced the nominative in 
the plural of the article ; in the singular, on the other hand, the 
nominative has replaced the accusative, of which the Middle 
Irish already affords examples (cf. der in the Allemannian dialect 
of German). The syntactical peculiarity of the Old Irish of 
putting the accusative in many instances in place of the nomina- 
tive, especially in the passive, and the complete similarity of 
both cases in the plural, which often originally existed or arose 
at an early period, as well as the slight difference in the singular 
masculine, which completely disappeared before tenuis, and s, 
/, facilitated this intermixture ; in addition to this, in the article, 
both were from the beginning alike in the feminine plural ; and 
in the noun, the accusative and vocative plural were the same, 
the latter being the only true accusative form, which is still pre- 
served, and which may also be recognized as such by the unal$ 
tered anlaut of a following adjective. The confusion has gone 
so far in the spoken language, that this form occurs for the dative 
in the plural even after prepositions, one says, indeed, do na 
fearaibh (to the men), but also do na capuil (to the horses), 
O'Donovan, 83 seq.—7rpbg tovq (ro7c) tWoi; O'Donovan directs 
the supposed accusative to be put after gan (without), and idir 
(between), in the singular, in reality, therefore, the nominative. 
The true accusative form is to be found, on the other hand, in 
the so-called dative singular, for o^nm-bdrd is as little a true 
dative as the French au poete ( = ad ilium poetam). Even in 
Old Irish the dative distinguished itself from the accusative in 
the vowel only in the a- (ia-) and w-stems, which were capable 
of an u umlaut, and this distinction must have ceased i^ Modern 

On Phonology in Irish. 155 

Irish with, the loss of this umlaut ; all feminine and .consonantal 
stems formed both cases alike, from the beginning with i-uni- 
laut; there only remained, consequently, the difference of the 
auslaut towards the anlauts following. But we have already 
seen in the case of the pronominal suffixes, that the dative has 
only maintained itself after de and do in diobh and doibh; on 
the other hand, the accusative has come in after all other pre- 
positions, as the peculiar phonetic phenomena in aca, etc., show; 
consequently in that which the grammarians call the dative sin- 
gular, a true dative is only to be recognized after de and do; 
after other prepositions, on the other hand, the accusative ; and 
we should not wonder that in Keating, and in the North Munster 
dialect, the article gives rise to aspiration only in these cases 
(both prepositions, except in the County of Kilkenny, soimd alike 
do), while everywhere else it produces eclipse. Here, also, then, 
similarity of form has gone hand in hand with syntactical cor- 
ruption ; the Modern Irish is surpassed in the latter respect by 
the Modern Greek, which has wholly lost the dative, and even 
combines cnrb and fd (/uleto) with the accusative. The occur- 
rence of the dative after all prepositions (even gan and idir) in 
the plural of substantives, is, no doubt, due to an effort to gain a 
prominent distinction, which was not given here by the form of 
the article (iia without change of sound). Hence there have 
been preserved pure in the written language, the dative plural, 
only that already in the earliest times the article had begun to 
become truncated to na, the genitive singular and plural, and 
the nominative singular, in all instances; the dative singular, 
on the other hand, only after de, do, the acccusative singular 
after the other prepositions (gan and idir excepted), never as 
objective case, the nominative and accusative plural only where 
they have sounded alike as in the vocalic feminine stems, other- 
wise they are always confounded. If we now study the treat- 
ment of the anlaut after the article, everything may be satisfac- 
torily ..explained conformably to the old rule, e. g.: — 

m. n. 

an t eari 

an fear 

an sruth 




an ein 

an fir 

an t-srotha 




dd'n ein 

do'n fear 

do'n t-srutli 




o'n ein 

o'n bh-fear 

o'n srufh 



SQ- D. 

an uair 

an chlann 

an t-suil 




na h-uaire 

na chloinne 

na sulci 




ddn uair 

do' 'n chloinn 

dd'n t-suil 




o'n uair 

o'n g-cloinn 

o'n sul 



pi. n. 

na h-uaire 

na clanna 

na sidle 




na n-uar 

na a -claim 

na sul 



do-na-li-uaraibh 6 na clannaibh do na suilibh 


156 EbeVs Celtic Studies. 

"When, therefore, s suffers the so-called eclipse after is (in) as 
in Keating — is in t-saoghal (in the world), the dative form is 
there exceptionally preserved, while 'san seanchus (in the history) 
contains the usual accusative form ; it is perfectly according to 
rule that t and cl should remain pure after do'n (den, isin) for 
they are withdrawn by the n from the aspiration which should 
occur here ; after other prepositions the strict rule requires, how- 
ever, eclipse as well as in the genitive plural. But even the 
most abnormal modes of treatment of anlauts (as in Kilkenny 
and Tipperary, where b, f, g, suffer eclipse, c and p aspiration 
after all prepositions, and s is eclipsed, that is, aspirated by t; a 
real dative is preserved here, as the aspiration of the c, p, s 
shows, but b and g are assimilated with the nasal to m and ng y 
the softening of the f to b h is curious) agree, however, in 
this, that t and d, after the article, are nowhere aspirated ; proof 
enough that in the Old Irish, also, we have to deal with a dis- 
tinct law, and not with a negligence of writing. If individual 
writers have also changed the anlaut of substantives without the 
article, e. g., have eclipsed in the genitive plural, it is because 
they have totally misapprehended the cause of the phenomena, 
it is, therefore, wholly unjustifiable. 

That the adjective after the article is subject to the same laws 
of anlaut as the noun substantive, may be concluded a priori; but 
in general the case occurs very seldom, as the adjective comes 
mostly after the substantive, in the opposite case composition 
takes place, although they are sometimes separately written, as in 
arnoib briathraib, Zeuss, 926 (read amoibbriatliraib, as the ab- 
sence of the ending requires). In Modern Irish such combina- 
tions are, to be sure, mostly written separate, but the adjective 
remains unchanged, and the anlaut of the substantive is aspirated 
(except in instances like seanduine) so that the composition is 
readily recognizable (O'Donovan, 347, 349). — Besides the pro- 
nouns each, cecli, nach, alaile, and the cardinal numerals which 
regularly precede (indala appears to be compound, the ordinary 
numerals besides cetne and tanaise always precede the substan- 
tive) sain occurs now and again inflected before the substan- 
tive (saini persin in the nom. fem. plur., but hipersonaib sainib), 
mostly, however, compounded ; title fluctuates, cStne, also, mostly 
precedes, but sometimes comes after the substantive ; aile and 
tanaise are, on the other hand, always placed after it. 

(2.) The same influence of the auslaut on the following anlaut 
occurs, of course, between adjective (adjectival numeral and pro- 
noun) and substantive, whenever the adjective has attributive 
value, whether it goes before or after ; the examples in Zeuss are, 
however, few, as the aspiration is never noticed in the case of b, d, 

On Phonology in Irish. 157 

g, m, and very irregularly in the case of s and /, and for the 
reasons above given must often be suppressed in the case of t 
(d). Examples for the aspiration of the adjective: in the nom. 
fern., mo thol cholntde, each thuare; in the gen. neut., indfolaid 
chetnai; in the dat. neut. isindanmaim chetntdiu, hi togarmim 
freendaire ; 121 in the dat. fern., dingutai thdisig, iar ridgoil chenelaig, 
for learn chli; in dual fern., diguttai fodlaidi. The aspiration 
is suppressed according to determinate rules in: gen. masc, in- 
chnunn toTrthich, toirthig ; in the dat. neut ;, far diidl tan. (tdn- 
aisiu); in the dat. fern., do persui tanaisi, hi perstn tanaisi, isin- 
depistil tdisich; from negligence in the dat., huandlwtthi sei'm, 
etc. Examples of the transvected nasal are given at p. 90. 
That the principle has not lost its force, we see in the Modern 
Irish, where we again find in the adjective placed after the sub- 
stantive aspiration and eclipse, under the same conditions (and 
also the same confusion in the dat. sing.) as in the case of the 
substantives; aspiration occurs in the voc. sing., only after con- 
sonants in the nom. plur., not after vowels, na fir threana, but 
treasa mora, because in the former -i was the original auslaut, 
and in the latter s; in like manner, the voc. plur. preserves the 

original anlaut — a fear a triana. 

Examples for the aspiration of the substantive: in the gen. 
masc, alaili thriiun (I have not as yet found nom. fern, sing.); 
in the dat. masc, re each thuisiul, dnach fochun ailiit, isindinchorp , 
in den sosiith sill-, 1017, neut., icachthir,fem., on chetni phersXn ; 
in the nom. plur., (inchamthuisil appears to be a compound), 
itchethir chet; the aspiration is prevented in the dative (f.) ondd- 
entoisrinn; left out from negligence in con alailiu fogoir, 6 din 
sil., don chetni per sin, hi cetni persin, in den sill., 1017, and re- 
markably enough in almost all cases which I have yet found, 
after each — gen., catch ceneuil, dat., do cech cenelu, do each eeneoll, 
do each ceneolu, ocech cenelu, hi each ceniul, do each ceniul, do each 
cathrur, do each corp, hi cech caingnim, icach sens (does a similar 
euphonic law rule here, except in the last instance, as in the case 
of t after dentals ? JYephplandatu maybe regular). The curious 
cachnden chrann, 999 (the subject in the ace like cech consain, 
1017), maybe explained as composition, as in the nom., denchoim- 
diu, oinchorp, 587, ace in denchorp, 580, tri den pheccad (on the 
other hand each den creitfess, gen. fern., inna oena meite), and is 
therefore to be read, cach-n denchrann; aon is always to be looked 
upon as in composition with its substantive in Modern Irish, for it 

121 These two formulas show that of the two attempted explanations given in 
Bsitrage, I., 451 (Stokes' " Observations on the Irish Declension") only the 
second is possible : anma?ibi, anmambi, anmammi, anmaimm. 

158 EbeTs Celtic Studies. 

aspirates tlie anlaut of the substantive, e.g., aoii cJiluas (one ear), 
£, aon chraun masc. If the masc. follows da (which in Mod. Irish 
has also replaced the fern, di), in the same form as in the nom. sing., 
the fern, in the same form as in the dative singular, but both being 
aspirated, da clirann, da chluais, and that the adjective in the 
plural occurs with aspiration, we have an exact correspondence 
with the little we know of the dual in Old Irish (see p. 86 seq.), 
and even the n of the old neutral form dan, is still to be recognized 
in the eclipse in da d-trian (two thirds) ; O'Molloy had also sus- 
pected relics of the dual in it, and ODono van's argument against 
this view, as well as in respect of the form after cead, mile and the 
decades, is only in part true ; that is, the apparent similarity of 
form which as a rule occurs between the nom. sing, and plur. of 
both genders, and between the nom. sing, and dual masc, has 
gradually led to the use of the nom. sing, after these numerals 
even where the gen. pi. (or nom. dual masc.) had preserved the 
original difference. The occurrence of eclipse after seacht, ocht, 
not, deich, is easily explained, and was already founded in the 
Old Irish by the n after these words. The explanation is more 
difficult of the pure anlaut after cidg and mile, where we should, 
certainly, have expected primitive vocalic auslaut ; in the former, 
according to the analogy of Tri/nirE, quinque, in the latter, in con- 
sequence of the Old Irish di mili, which points to a feminine ; 
the consonantal auslaut which we must assume in cead (perhaps 
already O. Ir. masc, compare the above cetliircMt) and se (for 
ses = sez), as in the plural forms, tri and ceithre, is easily under- 
stood; we have Osteins mfiche, and the remaining names of the 
decades, whose nom. sing, are pronounced fiche, tricha triocha, 
gen., fichet fichead, *tric1iat triochad, and nom. plur. (like the 
dat. and ace sing.) *fichit ficliid, triclitt trochaid, etc. 122 

(3.) The combination between the substantive and a succeed- 
ing genitive is much weaker. Examples of the nasal preserved in 
the accusative, (frislond nilfolad, 1029), nominative neuter (torbe 
nimdibi), and the genitive plural, are numerous enough (compare 
On the so-called prosthetic n, p. 90) ; but, as has been already 
above remarked, no very particularly close combination is neces- 
sary for this ; on the other hand, aspiration occurs rather seldom. 
The nom. fern., trebaire chollno, toll chollno, ciall chesto, ciall 
chesta (the neut., ahun thriuin is singular, beside the ainm-n of 
the examples in p. 91), dat., hitosiig sain, 1011, do immfolung 
fail, 1016, in den sosuth sill., 1017, do thcridbse superlaTt, 6 ilioil 

122 The doubt expressed at p. 433, vol. I., of the Beitrage is removed by this ; 
fich'e, tricha, for primitive i-icint, *tricant, now approach much closer to the 
Sanskrit trimqat, and are a mean between the latter and the Latin viginti, tri- 
ginta ; in meaning they express the Greek slicac, rpiaKccg. 

On Phonology in Irish. 159 

cholno, do lani chetbutho, are opposed, for example, in great num- 
ber, besides the regular ond des tudithe, dothabairt toirse, by the 
nom. fern., bandea cruithnechta, bandea tened, hii*es creitme, cidll 
cech miad (tir tairngiri may, like tir-n-erend, depend upon a 
change of gender) ; by the gen., eisseirgi cr., the dat., do hicc cdich, 
illestur ferce, fomdm pectho, a rainn pectlia, di red pectho, do de- 
chrugud persine, isdiri ceneoil, hi claar cridi, di muntir cessair, do 
each ceneolu serbe, oc ascnam tire tairngiri, hifoisite cesto, so that 
it would appear the fluctuation could scarcely have been here 
confined merely to writing. According to O'Donovan, 368, seq., 
aspiration in Modern Irish, also, is generally only usual in proper 
names, although Keating used it also in other cases ; but it has 
here inorganically extended itself to cases like Airdeasbog Chaisil 
(the Archbishop of Cashel), and consequently is used as a purely 
dynamic agent ; on the other hand, the above mentioned excep- 
tion, which the family names with and Mac make, rests fully 
upon the Old Irish anlaut laws. 

(4.) The pronouns stand in such intimate combination, as well 
with the substantive as with the verb, that many depend upon 
both parts of speech, not merely as enclitic or proclitic, but even 
penetrate between the preposition (verbal particle) and the verb. 
Thus the anlaut of the noun is under the influence of the so- 
called possessive pronouns, i.e., the genitive of the personal pro- 
nouns, whether the latter appear in their complete form (absoluta 
of Zeuss), or in a shortened form (infixa of Zeuss) ; mo, do, 3. 
masc. d end in vowels, o. fern, a., originally in s, the plurals, ark, 
farh or barn, i.e., bharh and ah, in nasals, hence : — motliol, moch- 
land, imchidmriug, domthoschid ; thual=:dofual,itchdimthecht (in- 
accurately, dosenmdthar, itsenmdthir, cutseitchi); achesta (inaccu- 
rately apectha); aggnim; arnet, arndiis, armbrethre, arloure-ni, 
arsoire-ni ; farnintliucht, farcluu,forserce ; ananman, ambes,accur- 
sagad. This influence is even now still felt in : — mo 'mil, mtfuil, 
do chos; a cheann; a ceann, a h-inghean; dr g-cinn; bhar g-cosa; 
a g-cinn. The anlaut of the verb is dependent upon the pre- 
ceding personal pronoun (infixa of Zeuss), but the decision as to 
their original auslaut is rendered more difficult by the contra- 
dictory ways in which they are written, and also by the circum- 
stance that Modern Irish has not this kind of combination. Vo- 
calic auslaut appears to be certain in 1. and 2. sing. — ni m 
charatsa, nomthachtar, nimtha, nimptha, foincl iridic! ifider - sa, 
nudamchrocha, cofordumthesid-se, fritumthiagar, fordomchomai- 
ther, rotchecldadar; we find, however, condumfel, aromfoimfea, 
immumforling,fomfirfidersa, romsoirsa, coatomsnassar ; rodchur- 
sach, to be sure, is found in 3. sing. ; but, on the other hand, we 
have immidforling , cenodfil, rondpromsom (with rel. n), n- and 

160 EbeVs Celtic Studies. 

s-, appear never to aspirate — ronsdir, nisfabnr, there is, however, 
nonchretid-si (ut in eum eredatis) ; we read in 1. plur., fonsegar, 
nunsluinfem-ni, nonsdirfea, ronsoirni cininfil, ronjitid-ni, ninfor- 
teit-ni (conintorgditar, homcntogaitarni, nintd are indifferent in 
consequence of the n-t), in 2. plur. atobci, nobcarad, fordubcech- 
na, forndobcanar, rondobcarsam-ni, robcar-si, nondubcairim-se, 
robclandad, nibtd, dobtromma, atobsegatsi, cotobsechfider, nondob- 
sommigetar, nobsoirfa-si, nachibfel, condibfell, manudubfeil, rob- 
fothiged, and jet nidan chumachtig and atdubelliub (i. e., atdud- 
felliub) appear to point to aspiration ; in 3. plur. da-, as well as 
sn-, seem to be without aspirating power — nodascara, rondasaib- 
set, nondasoirfea, noshguid-som, dosinbera, nisfttir, nosmoidet, nis- 
fitemmar, rospredach, roscomal. The combination between the 
verb and relative pronoun is equally close — an (anasbiursa, ar- 
rocar, acarthar, apredchimme) and no (nocretim, nopredchim-se, 
correctly or negligently written ?) ; it is curious that in Modern 
Irish the nom. a aspirates and eclipses only after prepositions, or 
as absolute neuter (what, all that). 

Enclitic pronouns and pronominal adverbs are, for their part, in 
respect to the anlaut, under the influence of the preceding word, 
the suffixed pronoun sa (se, so, su), even in respect to the vowel. 
We must, consequently, conclude from the circumstances that the 
s remains uninfected, that an original double consonant existed in 
anlaut, not only for som (sein) as Stokes has correctly remarked 
(Beitrage, I. 469), but also for su, sa, so (se, siu, sin), espe- 
cially as intitliall, with aspiration, stands opposed to intfsiu, and 
messe, tussu are found, whilst with the article we have, as was 
mentioned above, not intso or indso, but inse, insin; only it re- 
mains doubtful whether all these pronouns belong to the same 
stem (say sva-n), or whether the -sa after pronouns is to be sepa- 
rated from the (as it appears) adverbial -sa, -sin, after substan- 
tives. Simple anlaut, on the other hand, is betrayed by the dative 
siu, and the compound pronouns side (saide, nom. plur saidai, 
Z., 9), sodain by the aspiration in desiu, Z., 595, and 6'suidi, dat. 
fern., olsuide, am. sodain, arsodain, olsodain, olsodXn, fosodm. 

(5.) I have nothing to add to what Zeuss has said on the treat- 
ment of the anlaut after prepositions and other particles ; that 
the s after reh, con, in, iarh, for, tri (and in part also after la, a, 
fri) belongs to the article, is now, I believe, generally admitted. 
I know no explanation for the hardening of mediae in the verb 
substantive (also ni tenat, Z., 585, for ni denat) after to and 
ni, which, nevertheless, otherwise produce aspiration ; trithemel, 
trichretim, along with tresinfuil, among others, is equally striking. 
The dies (duus) before in- in the indirect interrogative is, no 
doubt, contracted from do-fius, du-Jkis (ad sciendum). Com- 
pare English to wit. 

On Phonology in Irish. 161 

(6.) The action of the verbs on the object as regards anlaut 
must have been already in Old Irish very weak. Zeuss gives 
only two examples, and O'Donovan has nothing about it in 
Modern Irish ; on the other hand, the aspiration • of the anlaut 
after ba, budh is given by him, also, as a rule, while in the docu- 
ments in Zeuss, more examples without aspiration after the root 
bu may be found than with it. 

§. 6. Loss of P in Celtic. 

One of the most interesting phonetic peculiarities of Gaedhelic 
is a certain aversion to p, which is manifested in different ways. 

Firstly, the Gaedhelic, as was long since remarked, has very 
frequently preserved the guttural where other languages, especially 
the Greek and the Kymric, have allowed the labial to replace it: 
thus, in accordance with the Latin, as opposed to the Greek and 
the Italic dialects, in the interrogative pronoun and all deriva- 
tives, Ir. cdch = Kymr. paup, quivis, Gaedh. nach wc/i = Kymr. 
nep aliquis; in the numerals Gaedh. ce£foV := Kymr. petguar 
four, and Gaedh. co/c = Kymr. pimp five; further, for example, 
in Gaedh. mace =Kymr. map films, Gaedh. eland = Kymr. plant 
proles, Gaedh. crarm = Kymr. pren arbor, Gaedh. cren = I£ymr. 
prenu emere, Gaedh. ech = Kymv. *ep> equus (Gaul, epo-, Welsh 
ebawl a foal), Gaedh. seek praeter = Kymr. hep sine, also, no 
doubt, Gaedh. cenn = KymT. penn caput, although Pictet (Bei- 
trage 86) considers penn =pinda older. Compare, also, Gaedh. 
sechim sequor, sechitir sequuntur, in opposition to which the 
defective Welsh heb inquit may be equated with the Greek 


Secondly, even c or ch has sometimes replaced primitive p xn 
even in loan-words, as caisc (pascha), cor cur (purpura) ;" 1 - 4 the 
circumstance is somewhat different with cuingeis, which, like 
O.H.G. fimfcliasti is only half borrowing, half imitation of pente- 
cost ; I cannot, however, look upon fescor or fescar vesper, as bor- 
rowed, for the Welsh ucher, as opposed to Corn, gwespter, Arm. 
goitsper, likewise betrays a guttural (ch=sc) like Lith. vakaras, 
Slav, vecerit. Cht is found for pt (as in Low German, nichte for 
nifte) in secht septem, seclitmaine septimana, necht neptis. 

But in anlaut an aversion to p shows itself in an especial 

123 To this category I also reckon the first guttural in cdic, as in Lat. quinque 
and coquo, which I attribute to assimilation (as in pare the second labial in iriince. 

and 7T67TWV). 

124 Pott, Hallesche Literarische Zeitung, 1844, S. 289, Anmerkung. 

162 EbeTs Celtic Studies. 

manner, not only in inconvenient combinations likeps, where, for 
instance, Gaedh. salm agrees with O.H.G. salmo for psalmo, but 
in the most convenient pi and pr, nay even before vowels, and not 
merely in Gaedhelic only, where perhaps the majority of cases 
of p- anlaut is due to borrowing (as in German, cf. peccad pecca- 
tum, persan persona, precept praecepturn, amprom improbus, prim 
primus, for the true Celtic cetne), but frequently also in Kymric, 
which is otherwise, however, as little averse to p as, perhaps, the 
Greek. It especially strikes one that, at first sight, we cannot 
discover, in both branches of the Celtic family, a single one of the 
many prepositions in Sanskrit and the other cognate languages 
withp- anlaut (para, pari, pra, prati, and their relatives). Pictet 
and Bopp have assumed that the p in these words has passed 
either into b or /, and very little of importance can be objected 
against the examples of the b for p in Pictet (De l'afnnite, etc., 
p. 49), isolated examples also occur in all languages of an irre- 
gular change between tenues and mediae, in Celtic, for example — 
Gaedh. gabar, Kymr. gafar=.\udX. caper, O. Norse hafr, A. Sax, 
heifer; Gaedh. gabdl = I£ymr. kafael, Lat. capere, Goth, hafjan, 
conversely, Gaedh. tenge = Goth, tuggo; Gaedh. ithim = Skr. admi, 
Lat. edo, Goth, ita; but the pretended change o£p into /is there- 
fore the more doubtful. Scarcely one of the examples quoted 
has direct evidence m its favour, but certainly the parallel 
Kymric, gu, gw, does not admit of the assumption of the direct 
passage of p into /, at most, of one through the mediation of v, 
from which the Gaedhelic f, and the Kymric gu,*\veYe then 
evolved according to their special phonetic laws. So, for example, 
Gaedh. frith, Kymr. gurth, certainly admit of being connected 
with the Sanskrit prati by a Celtic fundamental form *vrith, 
*verth ( = *vrati, *varti), by which the aspirates would be deve- 
loped in both languages perfectly according to rule, in the Gaed- 
helic between the vowels, in the Kymric in the position rt. The 
end vowel in isolated use must then, however, have dropped off 
very early, as the Gaedhelic has there only the form fri, which 
does not infect the following consonants ; for the o before the ar- 
ticle belongs as little to the preposition an in this case as in re, iar, 
in, tri. But the transformation of the Sanskrit pra to Gaedh. for, 
Kymr. guor, appears altogether improbable to me ; for the Celtic 
preposition (with which the intensive guor-, Gaul, ver- appears to 
be identical) is obviously related to Gaedh. fo, Kymr. guo, in form 
and meaning exactly like super to sub, virap to vtto, Goth, ufar 
to uf (which also agree in the double construction), therefore, 
also, as Skr. upari to upa. Only a doubt can, therefore, exist 
as to whether the Celtic had perhaps (like the Slavonian in na = 
ava and po = upa) dropped the initial vowel, and then changed 

On Phonology in Irish. 163 

p into v, or whether it had softened and suppressed the p after 
the u, so that the fundamental forms *va and *vari from u(p)a and 
n{p)ari, common to the Gaedhelic and Kymric, had developed 
themselves; the latter is my subjective conviction. If the i, 
dropped in for, guor, no longer exerts, almost anywhere, 125 an 
action upon the following consonant, it shows that the Celtic 
agrees with the Latin, Greek, and Gothic in the early rejection 
of that vowel ; but perhaps a trace of the i may be recognized 
(as in O. Norse yfir in opposition to Gothic ufar) in the Gaulish 
intensive prefix ver-, the e of which may have arisen either di« 
recibly, or through the intermediate stage of i from a by the in* 
fluence of i in auslaut. 

The Sanskrit pra and pari are rather to be found in a fourth 
class, among words which have wholly lost the p in anlaut, as 
in the Gaedhelic t'as^ = Kymr. pise, pysg piscis, athir pater, 
which includes in both languages the root Skr. par (pr), which 
always appears here, as in German, Greek, and Latin, with I for 
r. To this category belong, with a preceding liquid, Gaedh. Ian 
= Kymr. laun (Welsh llawn, Corn, len leun, Arm. leun) plenus, 
lane plenitudo, lanad and linad implere, rolin implevit, Welsh 
llewni implere, lloneit plenitudo (quantum implet) ; with a pre- 
ceding vowel, Gaedh. comalnad impletio, comalnadar implet, 
comalnamar implemus, comallnithe impletus. From the same 
root descends further *paru much = Skr. puru, Gr. iroXv, Goth. 
jilu, which the Gaedh. il (for pil) very accurately represents, 
whence ilar multitudo, ilde, pluralis; the Gaedh. comparative 
lia agrees with the Greek ttXhwv, Lat. plus; compare further 
Welsh liaus, lliaws, laws, multus, multitudo = Corn, luas, leas, 
W. llaiver = Corn. Hewer multus, Gaedh. laur, fowr=Corn. loar 
sufficiens, satis and loure sufficientia, W. lluossyd multitudo. 

Similarly Gaedh. lethan, Kymr. litan, llydan broad, Welsh 
lledann to spread out (llet) lied and llyd latitudo, connect them- 
selves with Skr. prthu for prathu, Gr. TrXarvg; the Kymric 
adjectives in -lit, -llyd, fern, -lied with the meaning " full of some- 
thing", if they are really compound, belong in their second part 
either to the root in question, or to the preceding one. 

I now likewise recognize the Skr. pra in the prefix ro, which 
appears in inseparable combination as an intensive particle, and 
in separable combination, as nota praeteriti especially; to the 

125 We find, nevertheless, in O. Ir., the secondary forms forchanim forchun 
praecipio, forchain praecipit, forthe'it adjuvat, forchongrim mando jubeo, for chon- 
gair mandat, along with Jorcanim (forcetal doctrina), forfeit, forcongrim for- 
congur, forcongair, in Modern Irish, foircheann for the old forcenn finis ; the rarer 
form Joir- owes its i, no doubt, to the influence of the vowel of the following 
syllable (as in foirbthe for forbuide). Zeuss, p. 212, also mentions eclipse in 
Kymric alor^ with aspiration, which likewise proves vocalic s-enesis. 

164 EleTs Celtic Studies. 

same stem belong Gaedh. re, ren, remi (superlative form as 
primum?) and Kymr. rac (=:Skr. prac?) I suspect the Skr. 
pari [but compare the next section] in the Gaul, are, whose 
fundamental form appears to be *'ari; compare Gaedh. ar, air, 
er, ir, Kymr. ar, er, yr, which may be very well compared with 
the Gr. irepi, in the meaning generally, and in the shades of 
meaning which it expresses ; thus the intensive er agrees with 
the Gr. wipi, 7T£p, Lat. per in permagnus. If a separation could 
be carried out between ar and air, 1 would prefer comparing ar 
with the Gr. irapa; Caesar's A rmorica might then be justified as 
TrapaXid; Aremorica (jrspi^aXaacnog) may, however, be also ex- 
plained. The Com. and Arm. war contains, perhaps, an indica- 
tion of the lost labial; the form am- also, w r hich the primitive 
an assumes in Gaedh. amires (unbelief), amiressach (unbelieving), 
may owe its origin to the subsequently dropped p of ir-es. 

Finally, the Gaedh. ire ulterior (erroneously described by 
Zeuss as a comparative, for ireiu is the comparative), may be 
referred to the stem Skr. para, and compared with its nearest 
akin the Greek Trepalog. 

§. 7. Loss of P in Celtic, continued. 

Since the preceding was written, I have found an interesting 
example of the loss of p in anlaut in in (&yis) = *ethn, V. hethen 
(volatile), W. 1, aetinet plur., (volucres), — with the derivatives V, 
idne (auceps), ydnic (pullus), the compositum, W. 1, etncoilhaam 
(angoror), — and the related words, W. 3, adaned, plur., (pennae), 
W. 1, atar, 2, 3, adar, (coll. aves), sing. W. \ f eterinn masc, 2, 3, 
ederyn (avis, volucris), evidently from the root pat (rrtrofjiai), — 
compare Skr., patatra, patra, A. Sax. fe*8er, O.H.G. fedara, Gr. 
TTTtpov, and Lat. penna, from *petma. Pictet (Beitr'age, II. 90), 
like Pott (Etymologische Forschungen, I. 2te Aim 1 ., 699 seq.), 
equates Gaul, are-, with the Vedic dra; I cannot, however, con- 
vince myself that this, in descent as in meaning, still very ambi- 
guous word has been preserved as a preposition in European lan- 
guages, 126 and therefore, I still assume the loss of a p in this pre- 
position; but I entirely give up the equation with pari, irzpi, to 
which I was even then persuaded with difficulty by the form 
air-. Many prepositions appear in Old Irish just as in Lithuanian 
(Schleicher's Lit. Gram., p. 133), in a double, nay, even in a 
treble form, a circumstance which I did not formerly observe, the 
shortest mostly occurring in independent use, the stronger in com- 

126 Also I do not see why (notwithstanding Pott's energetic protest against it) 
Lat. ad and ar, which only appear before labials, could not have coexisted 
during a long period as dialectically different forms, just as well as N.H.G. 
sanft and sacht, as the transition of d into r is proved by meridies. 

On Phonology in Irish. 16' 5 

position, and before pronominal suffixes, with which the pecu- 
liar intercalated syllables in Kymric may be compared. Exam- 
ples: in (ingiun, itossuch) in (inchosc) inaV (indiumm), con c. d. 
(condiuiti, coseitchi) com (comchesaa) rarely con (cosmil), cos 
(?) c. a. (co osnada) cue (cucci), reii (renairite, recach) rein 
(remib) remi (remiepur, remtliechias), iarh (iarmbaithius, iar 
timnu) iarma (iarmafoich, iarm(s)uidigthe) once iarh (iarfaigid) ; 
as (abas, asind-) ass (esib) as (asoirc), tars (tar crick, tar sin-, 
3. taraxs) tarmi (tairmthecht), tris (tritliemel? tresin-, 3. triit) 
trpmi (tremdirgedar), fri's (fricach, frissin-, 3. friss) frith' 
( frithcheist) seldomer fris (frisbiur); 6 (hotJml) uad (uad- 
fialichthi); even fortheit along with forfeit points to for. 
Thus ar (archiiinn, archenn) also represents undoubtedly a 
fundamental *ara (therefore, perhaps, *pard), air (airchinri) and 
airi (airiumm) on the other hand, very probably a strengthened 
form *are from *(jp)arai, and Ausonius' measurement Aremoricae 
need not be at all looked upon as forced by the hexameter ; hence 
*ara — ar is to trapa as *are = air is to trapai = Eith. pre (pry-, 
p?7-):=Slav.£>n, and Gaul, aremoricos would be *irapai6a\a<j<jiog, 
as the modern Breton arvorek TraoaOaXaaaiog. But the funda- 
mental meaning of ar- appears to be N.H. G. vor [Engl, fore] 
(jjro and prae) from which fur, with all its shades of meaning, 
was developed, which the English for and French pour might 
denote: ar ch iunn (vor dem angesicht, before the face), archenn 
(vor das angesicht), arse (pour cela, for that), arnaib uilib 
cumactib (prae omnibus potestatibus), airi (therefore, e'est 
pourquoi), doaurchanim (portendo), argur (Goth faurbiuda, 
N.H.G. verbiete), arah {pour que) ; even the conjunction ar 
is to be found in the English for; again anair is properly 
N.H.G. von vorn [Engl, from before], airther = 7rapoir£pog (irpo- 
TEpog), O.H.G. fordoro, aire and airecJi represent the Skr. purva 
and the N.H.G. vorig. Now how does all this agree with para, 
which the Gr. irapd is supposed to represent? At the risk of 
being considered a very great heretic in etymology, I answer, 
certainly not with Skr. para to which Greek and Latin forms in £ 
and e correspond (irapa either not at all, or only in certain combi- 
nations), and which is itself only weakened from *apard; but no 
doubt with the *pard of the primitive Indo-European language, 
which appears again in Skr. purd, but is preserved in Gr. rrapd, 
as is * paras, Skr. puras, in the Gr. Trapog ; for notwithstanding 
the Gothic faura, the Sanskrit purd has no more preserved the 
primitive vowel than p>ur as, because Goth, faur may be equated 
with it, hence not only Zend paoiirva, but also Old Persian pa- 
ruva represent Skr. purva, and the Goth, fairnja (fairneis) 
with the more modern derivative, represents the Skr. purdna, 

166 EbeTs Celtic Studies. 

all being relatives of the Lat. prae, pro, por-, trie Gr. irapa, 
wapai, irpo, tlie Lith. pra, pre, Slav, pra, pro, pri, as well as of 
the O. Ir. ar and air, and as vor ethically modifies its meaning 
to filr, so it also weakens itself in meaning to an (compare 
praebere = 7rape\£tv). I have already spoken above of Modern 
Irish ar, to which the supposition of Pictet respecting Old Irish 
for applies. 





[The following pages contain a translation of the part of the second chapter 
of the Grammatica Celtica of Zeuss, concerning the inflexions of the noun, to 
which reference is so frequently made in the Celtic studies of Dr. Ebel. One 
of the most remarkable features of Zeuss' work is the large number of examples 
taken from MSS. which he has brought forward as the basis upon which his 
grammatical canons are founded. Thus the examples given in the part of the 
chapter here translated fill considerably more than thirty pages. All these 
examples not being necessary for the purposes for which this translation was 
made, only a small selection of them has accordingly been given. 

(A) Declension. 
In the Old Irish language, the nouns of which have preserved a 
great variety of forms — hi this respect far surpassing the Welsh even 
of the same period — we find two orders of declension, of which the 
first, on account of the prevalence of vowels in the inflections, may be 
called the " vocalic", and the second, for a similar reason, the " con- 
sonantal order". To the former belong the adjectives, which do not, 
as in other languages such as the German and Sclavonic, possess pe- 
culiar forms of their own ; substantives alone are found in the latter, 
though in less number than in the first. In both orders the flexional 
vowels are either exterior, applied to the end of the word, or inte- 
rior, placed immediately before the final consonant, whether it be a 
radical or derivative one. There are, moreover, some anomalous 
nouns differing from the usual forms of declension, and exhibiting 
others peculiar to themselves. 

Substantives and adjectives of the masculine and neuter genders 
agree in their declensions. Those of the feminine gender have forms 
of their own. I shall give first a table of all the forms of declension, 
which I call series, with a paradigm of each ; and then substantives 
and adjectives from the codices confirming the forms of all the 
series here exhibited, or even such as present any slight varieties. 

declension of nouns Masculine and Neuter. 
Paradigms : I. — Cele (a companion). It has not appeared so neces- 

13 b 



sary to give an example of a derivative of this first series, such as 
echire (a horseman, a muleteer ?), tectire (an envoy), as of the follow- 
ing, on account of the internal vowels inflected : II. ball (a member), 
primitive, tuisel (a case), derivative example. III. bith (the world), 
primitive, dilgud (forgiveness), derivative. 

The neuter differs so far only from the masculine, that the ac- 
cusative and vocative are formed like the nominative ; and, in the 
plural number, the same three cases take peculiar inflexions, dif- 
ferent from the masculine, as will be rendered evident by the 
examples which follow : — 


























































declension of nouns — Feminine. 

Paradigms : IV. — tuare (food). V. rann (a part), primitive, briathar 
(a word), derivative. 



IV. Series. 













































I. Series. — Of nouns externally inflected, and ending in -e, 
which in the different cases becomes -i, -iu, -ib. Neuter nouns in the 
nom. ace. and voc. plural vary from -e to -i. 


Nominative. — Substantive Masculine — cele (a companion, husband), Wb. 
Sg. ; duine (a man), Wb. ; dalte (a disciple), etc. 

Subs. Neut. (I give examples only of such as are met with the article), 
anesseirge (the resurrection), Wb. 30 b ; atrede (trinitas), acetharde (four), Wb. 
cumachtae (power), Sg. 6 a . 

Zeuss on the Inflexions of Nouns. 171 

Adjectives. Masculine, ce'etnefer (first man), Wb. 7 b ; intathir nemde (the Hea- 
venly Father), Wb. 4 b ; derivative adj. in de, te, the, are of frequent occurrence. 

Adjectives. Neut. anuile (all), anuilese (all this), Wb. 16 b ; ni nuae hdo anatrabsin 
(this possession is not new to him), Ml. 17 b . 

Genitive. — Subst. Masc. corp induini (the man's body), Wb. 12 a . 

Subst. Neut. claar cridi (table of the heart), Wb. 15 a ; comalnad sosce'Ii (fulfil- 
ment of the Gospel). 

Adj. Mas. comalnad indhuili recto (fulfilment of all the law), Wb. 20 a . 

Adj. Neut. dinsid cetni diil (accusative of the first declension), Sg. 91 b . 

Dative. — -u occurs frequently instead of -iu. 

Subst. Masc. do duiniu (to the man), Ml. 20 d ; donduini (to the man), Wb. 4 b . 

Subst. Neut. dondediusin (to these two), Wb. 9 C ; hi farcridiu (in your heart), 
Wb. 5 d ; In esseirgu, in heseirgiu (in resurrection), Wb. 4 b 13 b ; iarnesseirgiu 
(after resurrection), Wb. 3 C . 

Adj. Masc. donchoimdid nemdu (to the Heavenly Lord), Wb. 27 e . 

Adj. Neut.^ar cetnu diull (in the first declension), Sg. 90 b . 

Accusative. — Subst. Masc. imfolngi induine firian. imfolngi induine sldn (facit 
hominem justum, salvum), Wb. i d . 

Subst. Neut. ni dilgaid anancride (you forgive not the spite), Wb. 9 C ; pred- 
chimmi sosce'le (we preach the Gospel), Wb. 14 c . 

Adj. Masc. lasinnathir nemde (with the Heavenly Father), Wb. 19 d . 

Adj. Neut. cen imdibe stdride (without bodily circumcision), Wb. 2 d . 

Vocative. — Subst. and Adj. Mas. a iudidi{0 Jew !), Wb. l d ; a mar thormachtai 
(gl. macte, magis aucte) Sg. 76 a . 


Nominative. — Subst. Masc. comarpi (co-heirs), Wb. 19 c . 

Subst. Neut. -e in Nom. and Ace, ataat ilehenele (there are many kinds), 
Wb. 12 d . 

Adj. Masc. d€nemdai (heavenly gods), Sg. 39 a . 

Adj. neut., na accobra colnidi (the carnal desires), Wb. 20 c . 

Genitive. — budid innam railed talmande (victory of the worldly soldiers), Wb. 
ll a . 

Dative. — donah huilib doinib (to all men), Sg. 189 b . 

Accusative. — Subs. Masc. friar ceiliu (against our companions ; i. e. against 
others), Wb. 33 b ; eter doini (amongst men), Wb. 28 b . 

Subst. Neut. same as Nom. ; ruchualatar ilbelre (they heard many tongues), 
Wb. 12 d . 

Adj. Masc. farnuili baullu (all your limbs), Wb. 3 b . 

Adj. Neut. na huli dorigniussa (all that I have done), Wb. 24 b . 

Vocative. — No instances occur for this series in the MSS. Elsewhere, how- 
ever, the Voc. plural agrees with the Ace. ; and here it may be fixed for the 
masc. -iu, and for the neut. -e, -i. 

II. Series. — Internal inflection, whereby in several cases, especially 
the Gen. Dat. sing, and Nom. phiral, the signs of the cases — i and u — 
either accompany or suppress the final radical or derivative vowel. The 
vowels which are most frequently so affected are a and e. A in those 
cases either becomes ai (pi, u%) and au, or disappearing leaves the i 
and u. But e with i and u becomes i and iu. The vowels o, 6, d, of 
more rare occurrence, and sometimes a in position, never admit of u 
by their side, but with i they become oi (ui) 6i, at ; <?, for which eu is 
sometimes found, with i becomes etui, iui, eoi ; with u iu ; 6i and di 
are nowhere changed. Substantives and adjectives neuter take a 
in the nom. ace. voc. plural. 


Nominative. — Subs. Masc. inball (the limb), Wb. 12 b ; inmacc (the son), Wb. 
Sg. infer (the man), passim. 

172 Appendix. 

Subst. Neut. anaccobor (the will, desire), Wb. 3 d ; anderbad (the certainty), 
Sg. 90 a . 

Adj. Masc. inspirut noib (the Holy Ghost), Wb. 4 a ; derivatives in ach, ech are 
very frequent. 

Adj. Neut. atir romanach (the Eoman land). 

Genitive Subs. Masc. ainm thriuin (a hero's name), Sg. 96 a ; di muntir 

Cessair (of the family of Cassar), Wb. 24 b . 

Subst. Neut. imchloud diill (change of declension), Sg. 31 b ; recht haicnid (law 
of nature), Sg. 217 b . 

Adj. Masc. isinanmim inspiruto noib (in the name of the Holy Ghost), Wb. 9 C . 

Adj. Neut. asainreth indanma dilis (that is peculiar to a proper name). 

Dative. — Subs. Masc. dondaum (to the oxj, Wb. 10 d ; dofiur, donfiur, do Sen 
fiur (to the man, to one man), Wb. 10 b , ll c , 21 a . 

Subst. Neut. far cetnu diull (in the first declension; diall), Sg. 90 b . 

Adj. Masc. on spirut noib (from the Holy Ghost), Wb. 14 c . Adjectives in ach 
are not changed : donbrdthir hiressach (to the faithful brother), Wb. 10 b . 

Adj. Neut. ar anmmaimm dilius (for a proper name), Sg. 27 a . 

Accusative. — Subst. and Adj. Masc. ar 6en fer (for one man), Wb. 4 b . 

Subst. and Adj. Neut. ataidlech (the satisfaction), Ml. 23 a ; cen sdithar (without 
labour), Wb. 27 b . 

Vocative.— a fir (oh man !), Wb. 10 a . 


Nominative. — Subst. Mas. adimmaicc (you are sons), Wb. 9 a . Itcorp 
inboillsin (these limbs are a body), Wb. 3 b ; 

e is changed, as in gen. sing. : asberat mo beiuil (my lips say), Wb. 12 d . 

Adj. Masc. slain (saved, slan), Wb. 28 b ; adib iressich (you are faithful), Wb. 12 d . 

Subst. and Adj. Neut. differ by the termination a : dtercitla (their prophecies ; 
tercital) Ml. 19 b . 

Adj. Neut. cecha dethidnea domundi (all worldly cares), Wb. 3 d . 

Final i is also met, especially in derivatives : itsaini inna rinn (there are dif- 
ferent stars), Ml. ; isli (gl. sunk, stars) Cr. 18 b ; isli doibsom infechtsa innahi rup- 
tar ardda dunnai (those [stars] are now low for them, which were high for us), 
Cr. 18 b . 

Genitive. — Subst. Masc. irchre flatho romdn (the decline of the Eoman 
Empire), Wb. 26 a . Riagoil sengrec ([the] rule of the old Greeks ), Sg. l a . 

Subst. Neut. : airitiu na forcetalsin (the reception of these doctrines), Wb. 16 a . 

Adj. Masc. esseirge innanuile marb (the resurrection of all the dead), Wb. 13 d ; 
indocbdlinna noib innim (the glory of the saints in Heaven), Wb. 13 c . 

Adj. Neut. foragab duaidinna anman adiecta cen tabairt anman trenfriu (David 
assigned to them nouns adjective, without the addition of appellatives), Ml. 30 a . 

Dative. — Donaib ballaib ailib (to the other members), Wb. 12 b . 

Accusative. — Subst. Masc. far nuili baullu (all your members), Wb. 3 b . 

Adj. Masc. la marbu (with the dead), Wb. 25 b . 

Subst. and Adj. Neut., same as in the nom. fodaimimse imnetha (I suffer tribu- 
lations), Wb. 23 b . 

Vocative. — Subst. and Adj. Masc. a Romanu (oh Romans!), Sig. 41 b ; a 
Galatu burpu (oh foolish Galatians), Wb. 19 b ; a Judeu et geinti hireschu (oh 
Jews and faithful Gentiles), Wb. 3 a . 

Adj. Neut. inna anman adiecta (the nouns adjective), Ml. 30 a . 

III. Series. — Of nouns externally inflected, except the dat. sing., in 
which the internal u occasionally appears. Endings peculiar to this 
series, besides the u just mentioned, ib dat. plural, and u ace. and voc. 
plural, are : -o gen. sing, for which a is of frequent, and e of rare 
occurrence ; -a nom. pi., for which -e and -i are also met with ; -e 
gen. pi. : neuter substantives do not take an ending in those cases 
of the plural which differ from the masculine, but present their 

Zeuss on the Inflexions of Nouns. 173 

naked form. I have met with no adjectives of this series, unless it 
happens that tualang, pi. tuailnge (gnari) 127 , be one, Wb. 17 b . 


Nominative. — Sub. Masc. bith (the world) ; mug (a slave), Wb. And deri- 
vatives in as, chas, ad, id, thid, ud, igud. 

Subst. Neut. atir (the Earth), Sg. 33 a . 

Genitive — Masc, imnetha inbetho (tribulations of the world), Wb. 14 b ; mdrad 
daggnimo (magnifying of a good deed), Wb. 6 a . 

Neut., ainm renda (name of a constellation), Sg. 73 a . , 

Dative. — Masc, isinbiuthso (in this world), Wb. 12 d ; do mdrad doe (to the 
magnifying of God), Wb. 15 c . 

Neut., di thir (of the Earth), Wb. 9 b . 

Accusative. — Masc, tri denpheccad (through one sin), Wb. 3 a . 

Neut., crenas tiir (who purchases land), Wb. 29 d . 

Vocative. — I do not know an example of the vocative of this series. 


Nominative. — Masc adib mogce (you are slaves), mogi sidi uili (these are all 
slaves), Wb. 3 b . 7 d . The ending i is only found in sub. masc in -id, -thid : 
foglimthidi (disciples) 13 a . 

Neut., itsaini inna rinn (there are different stars), Ml. 

Genitive. — log apecthe (the reward of their sins), Wb. l e . 

Dative — diamogaib (to his slaves), Wb. 22 d . 

Accusative. — Mas., na ddnu diadi (the divine gifts), Wb. 28 c . 

Neut., inna mind (gl. insignia, celebramus nostras redemtionis), Cr. 41 c . 

Vocative. — Does not occur ; by analogy, bithu, gnimu, etc. 

IV. Series. — Of nouns fern, externally inflected, ending in -e and 
-i, and, therefore, corresponding to mas. and neut. nouns of the first 
series in -e, -i, and -u. 


Nominative. — Masc. : lane, lanoz (fulness), Wb. 26 d , 27 a ; firinne, (truth) 
Wb. 2 d . 
Adj. Jirinne rectide (righteousness of the law) Wb. 24 a . 
Genitive. — Maicc soilse (sons of light), Wb. 25 c . 
Adj. hifoirciunn na cetnae rainne (at the end of the first part), Sg. 18 b . 
Dative. — Subst. cofailti (with joy), Wb. 24 b . 

Adj. icomairbirt nuidi (in understanding the [New Testament]), Wb. 3 C . 
Accusative. — Subst.: cenfirinni (without truth), Wb. 2 a . 
Adj. tresinfuil spirtaldi (through the spiritual blood), Wb. 20 d . 


Nominative. — Subst. cit sochudi (though there be many), Wb. 4 d . 

Adj. inna ranna aili (the other parts), Sg. 22 a . 

Genitive. — Subst. do airbirt biuth inna tuaresin (to enjoy this food), Wb. 10 c . 

Adj. etarcne nardun diade (knowledge of the divine mysteries), Wb. 26 c . 

Dative. — Ibartolaib \_Inbartolaib c ?~\ marbdib (in your mortal wills), Wb. 3 b 

Accusative Subst. inna lobri (the infirmities), Wb. 6 C . 

Adj. adciamni na runa diadi (we perceive the divine mysteries), Wb. 12 c . 

V. Series. — Of nouns fern, inflected both externally and internally, 
and corresponding at once to Series II. and III. mas. and neut. Special 
vowel endings are : -e in gen. sing., -a in nom. and ace. pi. ; besides 
internal -i in dat. and ace. sing., if the last syllable admit of the in- 

127 [Tualaing properly means able, competent.] 

174 Appendix. 


Nominative. — Subst. (of frequent occurrence inthe codices): ess, iress (faith), 
nem (Heaven), lam (the hand), etc. 

Adj. also numerous : sere mdr (great love). 

Genitive. — Subst. tuag nime (rainbow), Sg., 107 b . 

Adj., airde serce more insin (this is a sign of great love), Wb. 24 c . 

Instead of -e, the regular case-ending, -o and -a occur (or vice versa -e for -o, 
-a in Series III. mas. and neut.), whether by affinity or dialectical variety ; luct 
inna cecolsa (those who are of the church), Wb. 12 b . 

Dative. — Subst., isindinducbdilsin (in this glory), Wb. 4 C ; isinbliadinsin (in 
this year), Cr. 32 b . 

Adj., o laim deiss (on the right hand), Sg. 17 b . 

Accusative. — Subst., tri kiris (through faith), W"b. 2 C ; pridchossa hiris (I 
preached the faith), Wb. 7 b ,yW toil de (against the will of God), Wb. 4 C \fri etdil 
(against Italy), Wb. 6 d . 

Adj., isarnach nindocbdil mdir (it is for every great glory), Wb. 23 b . 

Vocative. — A ndib ingen (oh holy virgin! gl. marg.), Sg. 112 a . 


Nominative. — Subst., lama et cossa (hands and feet), Wb. 12 b ; na bretha (the 
(judgments), Wb. 17 b , inna ranna (the parts), Sg. 22% 26 b ; na briathrasa (these 
words), Wb. 28 c . -e and -i also occur in many, as the result of assimilation : 
octhdelbce andsom (gl. sunt formce octo), Sg. 166 a ; na litre (the letters), Sg. 10 a ; 
inbertar epistli udin (shall the letters be sent from us ?), Wb. 15 a ; aihissi (gl. con- 
flictiones ; sing. nom. aithiss, Wb. 13 b , compos, ut iress ?) Wb. 29 b ; teora bliadni 
(three years), Cr. 32 b . 

Adj. in -a: beisti olca (evil monsters, or reptiles), Wb. 31 b . Adj. in -i: itne- 
phchumscaichti na teora litreso (these three letters are unchangeable), Sg. 10 a . 

Genitive. — Subst., etarcne naruun (knowledge of the mysteries), Wb. 26 c . 

Adj., inna teora liter (©f the three letters) Sg. 

Dative. — Ho Idmaib (from hands), Wb. 9 a ; donaib teoraib personaib uathataib 
(of the three persons singular), Sg. 186 a . 

Accusative. — Subst., adciamni na runa (we perceive the mysteries), Wb. 12 c ; 
fri tola inbetho (against worldly desires) Wb. 29 a . 

Adj., nigette [nigente ?] na brithemnachta becca (you would not form slight 
judgments;, Wb. 9 C . 

Subst. and Adj. in -i: acosmiligmer dvli ecsamli(we compare things dissimilar) 
Sg. 211 a . 

Vocative. — ni riccim forless a chossa (I require not your aid, oh feet !), 
Wb. 12 a . 


Nouns of this class end for the most part in consonants, or rather 
have in some cases consonantal endings which, being originally, no 
doubt, derivative, show traces of an internal derivative inflection, 
with the mutable vowels a, e, and i preceding the final consonant. 
The final consonants are the liquids in, n, r, and the mutes d, ch, 
which with the internal vowels form a series of terminations — ir, ar, 
ir; in, an, in; id, ad, id, etc. The one series of the substantives in 
-m and -im, which I place first, developes certain special forms. If t 
appears instead of a, two divisions arise : (a) an, in, ad, id; (b) en, 
in, ed, id. Which discrepancy of vowels can scarcely be ascribed to 
assimilation, in the face of such forms as senman, menrnan, foirbthetad^ 
orpamin, and others. 

Examples of the liquid series (I.) (II.) (III.) : ainm (a name), beim 
(a stroke), menme (the mind), ditu ditiu (a roof), athir (a father). 

Examples of the mute series (IV.) (V.) druid (aDruid), cathir (a town). 

Zeuss on the Inflexions of Nouns. 


I. Series. 

•V- . 

II. Series. 

III. Series 





' ^ 

Sing. Nom. 
























Plur. Nom. 

















i athrib 








IV. Series. 

V. Series. 


r 1 

druid fili 



druad filed 



druid filid 



druid filid 




druid filid 



druad filed 



druidib filidib 



druida fileda 


I. Series consists of some substantives in im, m, taking in the gen. 
sing, -a or -e; in the dat. -im, with duplicated m ; and in the plural 
either an or en, these two endings forming two distinct classes. In the 
first (a), the noun ainm, of constant occurrence, is proved to be of the 
neut. gender, from the passage (Sg. 56 b ) : ashdirruidig. anainmsin V2% (this 
noun is derived). Of the same gender, no doubt, are all other nouns 
of this form. Of the second class (b) but few examples occur, and these 
not uniform. There is no instance of a vocative in this or any of 
the other series. 


Nom. — (a) ainm, ainmm (a name), Wb. Sg. passim. 

(b) beim (a blow), ingreim (persecution), Wb. 18 d . 

Gen — (a) indanma dilis (of the proper name), Sg. 26 b ; (b) no example found 
in the codices. 

Dat. (a) isinanmim inchoimded ihu. cr. (in the name of the Lord Jesus Christ), 
Wb. 9 C . 

(b) ocmingraimmaimse (at my persecution), Ml. 33 a . 

Ace (a) cen ainm (without a name), Sg. 21 l a . 

(b) ni agathar dingreim (his persecution is not dreaded), Wb. l a . 


Nom. — (a) asbertar ananman (their names are mentioned), Wb. 28 a . 
(b) bemen digle (strokes of revenge), Wb. 17 d . 
GrEN.=r(a) diall nanmann (declension of nouns), Sg. 27 a . 
(b)foditiu nan ingremmen (endurance of the persecutions), Wb. 23 c . 
Dat. — (2L),inanmanaib hit. (in Latin names), Sg. 6 a ; (b) no example known; 
b&nnib in the table is, therefore, hypothetical. 
Ace. — (a) tre anman (by nouns), Sg. 29 a . 

II. Series. — Consists of nouns taking in the oblique cases an, in, and 
in, en, whence two divisions. To the first belong derivatives in -min, 
■man, -mn (which is reduced, however, in the nominative to -me, or • m 

128 [Uncontracted form ashdirruidigthe anainmsin.~\ 

176 Appendix. 

only), to the second belong derivative nouns in -in which in the same 
manner in the nom. becomes -iu, -u. In the oblique cases singular, 
likewise, especially the dative, other curtailed forms are found by the 
side of the fuller. These fuller forms of derivatives appear in the 
case of secondary derivatives : menmnihi (gl. dissensiones, from the 
sing, menmniche ; menme), Wb. 18 a ; brithemnacht, brithemnact (judge- 
ship), Wb. 6 b ; brithemandu (gl. judiciali, from the nom. brithemande — 
brithem), Ml. 26 c ; anmande (pertaining to the soul — anim), Wb. 13 d ; 
talmande (pertaining to the earth — talam), Wb. 3 d ; noidenacht (in- 
fancy — noidiu, an infant), Wb. 24 d ; caintoimtenach (well-thinking — 
toimtiu), Ml. 31 b ; ermitnech (gl. reverens — ermitiu), Ml. 32 b . For the 
vowels a, e, I add brdtharde, brotherly, from brdthir. 

To the second division (b) of this series belong numerous feminine 
nouns in tu, derived from verbs (tu for tiu, not to be confounded 
with masculines in -tu, gen. -tad, of the fourth series, and derived 
from adjectives). There are other feminines of the second class in 
-tiu, and in siu, derived also from verbs. In the first division are 
met both masculines, as, brithem, and feminines, as, talam, anim. 


Nom. — (a) isbeo indanim (the soul is living) Wb. 4 a . 
(b) toimtiu (supposition), Wb. 23 a . 
Gen. — (a) roscfornanme (eye of your soul), Wb. 21 a . 
(b) dliged remcaissen, dliged remdeicsen, (law of Providence), Ml. 19 d , 27 d . 
I) at. — (a) bum et talam, bum et italam (in Heaven and Earth), Wb. 21 a . 
(b) oc tuiste duile (at the creation of the elements, i.e., of the world), Wb. o c . 
Ace. — (a) accobor lammenmuin (desire in the mind), Wb. 3 d . 
(b) nertid arfrescsinni (he strengthens our hope), Wb. 5 d . 
The final iu, u of the nom. seems to have disappeared from some nouns in t, 
as, fortacht (help) Ml. l a ; bendacht (benediction), Sg. 


Nom. — (a) matuhe ata horpamin (if these be heirs), Wb. 2 C . 

(6) derbaishdisin (the very pronunciations), Sg. 3 b . 

Gen. — (a) do ice anman sochiride (for the salvation of many souls), Wb. 24 d . 

(6) dedliguth innan iltoimddensin (in right of these many opinions), Sg. 26 b . 

Dat. — (a) diarnanmanaib (for our souls), Wb. 24 d . 

(b) huafoisitnib (from confessions), Sg. 33 a . 

Ace. — (a) aforcital iccas corpu et anmana (the doctrine "which heals bodies and 
souls), Wb. 30 d . 

(6) for genitne (by genitives), Sg. 45 a . 

III. Series. — Of nouns of relationship, mas. and fern, in -ir, there 
is but one class, as e never occurs for a in the interior. 


Nom. — Athir (father), mathir (mother), brdthir (brother), Wb. Sg. passim. 

Gen. — Brdthir athar (gl. father's brother), Sg. 56 a . 

Dat.— Dondathir (to the father), Wb. 13 b . 

Ace. — Lasinnathir nemde (with the Heavenly Father), Wb. 19 d . 


Nom. — No instances in the codices, athir by analogy. 

Gen. — JIaic iridegaid anathre (sons after their fathers), Wb. 30 b . 

Dat. — Uambraithrib (from their brothers), Wb. 33 d . 

Ace. — Does not occur. I supply mas. athru, brdthru — fern, mdthra. 

Zeuss on the Inflexions of Nouns. 177 

IV. Series. — Of derivatives in -id, forming in the oblique cases by 
the variation of the internal vowels two divisions (a) ad, id; (b) ed, id. 
To the first belong very frequent nouns in -u, shortened from -id, as 
above, -u, -iu, from -in. The ending id, has been preserved only in 
the word druid, in the others becoming -e, as : tenge (a tongue), ume 
(brass). The terminations of the second class have also become in 
the nom. -iu, -i, or -e. The full form of the derivatives here also, as 
in the second series, is apparent from nouns and adj. of secondary 
derivation: filedacht (poetry; fill, gen. filed), Sg. 213 a ; oigedacht 
(hospitality, ogi) Wb. 26 b ; to which I add, Tenedon (tene, tened), a 
Gaulish topographical name. Further traigthech (gl. pedester ; trai- 
gid, Wb.) Sg. 38 b , 50 b . 

The nouns of both divisions are masculine. 


Nom. — (a). Abstract Nouns in u from adjectives are very frequent. The end- 
ing is either -u simple, or the fuller -atu, -etu. 

Adj. of different form taking -u: artu (height), = arddu, ardu (from art, 
ardd, ard); domnu (depth, from domun) Incant. Sg. So also -atu, -etu: ddnatu 
daring) Sg. 90 a . 

Adj. in -ide, -de, -te, taking -u : Sentu (unity, adj. Sente, Sende, Wb. 7 C ); corpdu 
(corporality, adj., corpde), Wb. So also, -atu, -etu : fiiuchaidatu (humidity, 
adj., fiiuchaide), Cr. 18 c ; foirbthetw (firmness), Wb. passim. 

(6) coimdiu (Lord), Wb. ; tene (fire), Sg., 69 b . 

Gen. — (a) tech nebmarbtath (house of immortality), Wb. ] 5 C . 

(b) bandea tened (goddess of fire, Vesta), Sg. 53 a . 

Dat. — (a) ondnephpiandatu (from the impunity), Ml. 28 a . 

(6) dofilid (to a poet), Sg, 14 a . 

Ace. — (a) cen torbatid (without utility), Wb. 12 d . 

(b) lassincoimdid (with the Lord), Wb. 25 b . 


Nom. — (a) doriyensat druid (druids made), Wb. 26 a . 
(b) intan labratar indfilid (when the poets speak), Sg. 162 a . 
Gen. — (a) from the Irish Annals : Muiredac na tengad (Muiredach [professor] 
of the languages) Tigern. ap. O'Con. 2, 275. 
(6) dolbud filed (poetic fiction), Sg. 71 b . 
Dat. — sechdaptkib (to the agents), W b . 19 d . 
Ace. — (a) lasnafiledasin (with these poets), Sg. 63 b . 

V, Series. — Of certain feminine nouns in -r, to which are added the 
suffixes -ach, -ich, -ig. The cases, though not all, of the noun cathir 
(a town), are met with in the codices, and the same declension is 
followed by nathir (a serpent) with the article in Sg. : indnathirsin (gl. 
natrix, i. e. serpens hie) 69 a . and doubtlessly by others in ir. Vestiges 
of this formation appear to have been preserved in the modern 
Irish: caora (a sheep, old form: cdir, cderf) Gen. eaorach, pi. nom. 
caoirigh. gen. eaorach, dat. caorchaibh, voc. (ace.) caorcha. It is cer- 
tainly preserved in some others in ir, as : lair (Old Irish lair, a mare, 
Sg. 49 b =Za-2r), lasair (a flame), gen. larach, lasrach, pi. laracha, las- 
racha. Here, also, the derivative ch, appears in the adj. cdirchuide, 
Sg. 37 (ovine) ; compare the Gaulish name Caeracates in Tacitus, 
and perhaps also Caracalla, the name of a Gaulish robe, (for caera- 
callaf), it is wanting, however, in trechatharde (gl. tripolites), Sg. 38 b . 

178 Appendix. 


Nom. — Cr. dim [din~\ issi inchathir (therefore Christ himself is the city), 
Wb. 21 c . 

Gen. — aitribtheid inna cathrach asb. tibur (gl. Tiburs : an inhabitant of the 
town which is called Tibur), Sg. 124^. 

Dat. — One would expect -ich, -ig, by analogy, but the contracted form of the 
nom. obtains in Wb. 13 b . : robot issinchaithir (he was in the city). 

Ace. — Romuil doforsat incathraig (Romulus founded the city), Sg. 31 b . 


Nom. — ilehatkraig (many cities); Sg. 13 a . 

The other cases must be supplied : Gen. cathrach. Dat. cathrichib (or cathrib?) 
Ace. and Voc. cathracha. 

Hie Dual Number. 

After the twofold formation of the Irish declension, we may here 
add a few words concerning this number, on account of the small 
number of examples furnished by the codices for all the series given 
above. It does not, of itself, denote two persons or things, as for 
instance in Greek, but constructed with the numerals dd, di, dib, it 
presents in the language of our codices mixed sing, and pi. forms, 
relics no doubt of more ancient forms peculiar to this number. 

The only form of the article in any case or gender, is, in before d, 
the initial letter of the numeral, which in one of the following 
examples is written dd, hard. 

We shall give, first, paradigms of the series of the first order, and 
then such examples as occur in the codices. The forms enclosed in 
brackets are hypothetical, or formed by analogy. 


I. Series. 

II. Series. 

III. Series, 

Nom. cele (i?) 



Gen. celi 

(baill) 129 


Dat. celib 



Ace. cele 






V. Series. 














Nom. — The Nom. Masc. appears to occur in the adj. dadruith cegeptacdi (two 
Egyptian Druids) Wb. 30 c . 

Neut. indagne (the two forms), Sg. 168 a . 
Gen. and Dat. — Gen. and dat. are not met. 

129 [xxxv. Eecte ball, which aspirates,* must, therefore, have had a vocalic 
auslaut (-o -au?) and so cannot possibly be (as Ebel supposes, On Decl. in 
Irish, §. 10 On the Celtic Dual, p. 85) identical with the gen. plur.] 

* We say (e.g.), aihair an dd maefhionn (fattier of the two fair sons), cailleach an dd adharc 
fhionn (hag of the tvro white horns). 

Zeuss on the Inflexions of Nouns. 179 

Ace. Masc. or Neut. : dobir dasale. dabir imduda are {arcaX Xsyopsva) 

Incant. Sg. !3 ° 


Nom. — Masc. : da mod, (two moods) Sg. 138 b . 

Neut. : comescatar da cenel indib (gl. two genders are mixed up in them), Sg. 61 a . 
Gen.— Of the gen. no instances. 

Dat. — Neut. : frisgair intestiminse dohdib dligedib remeperthib (this testa- 
ment answers to the two previous laws), Sg. 193 b . 
Ace. — imbir indamer (ply the two fingers), Incant. Sg. 


Nom. — Met da atarcud and (there will be two relations there), Sg. 198 b . 

Gen Cechtar da lino (either of the two parts), Sg, 162 b . 

Dat. — Corns. 6 dib nogaib (composed of two parts), Sg. 98 a . 

Ace. — Andiall foadanog (the declension in both its parts), 98 a . Sg. 

Neut. : indd €rrend (gl. stigmata, porto), Wb. 20 d . 

IV. series. 

Nom. — It dlgutai bite indeog. (there are two vowels in a diphthong), Sg. 18 a . 
Gen. — Fogor dagutce indeog. (the sound of two vowels in a diphthong), Sg. 18 a . 
Dat. — Evidently do dib guttib. 131 
Ace.— Adj. in Sg. 74 b , indi rainn ihgraidi (into two intelligible parts). 

v. series. 

Nom T>i huair (two hours), Cr. 31 b . 

Gen — Cechtar indarann (either of the two parts), Sg. 74 b . 
Dat. — Ni chen dliged anephdiall 6 dib rannaib (gl. alteruter, alterutrius non 
absque ratione non declinatur ; i. e. non declinatur e duabus partibus), Sg. 75 a . 
Ace. — Coitchenaso etir di drim (common to two numbers), Sg. 72 a . 

Duals of the second order are very rare. The following are in- 
stances : — 

Tuicsom inda nainmso (he understands these two names), "Wb. 21 d ; da druith 
cegeptacdi (two ^Egyptian Druids), Wb. 3(K 

Anomalous Substantives. 

Which do not follow a fixed rule and form like all those above 
enumerated, but have peculiar and shifting forms of their own. 
Of this kind are : dia (God), dia (a day), duine (a man), ben (a 
woman), rig (a king), Id (a day). 

I. Dia (God) : sing. gen. etargne hdee (knowledge of God), Wb. 21 a ; dat. 
6 dia (from God) ; ace. fri dia (with God), Wb. 20 d ; voc. a dde (oh God). 
Wb. 5 b ; plur. nom. de nemdai son (Heavenly Gods), Sg. 39 a ; dat. do deib (to the 
Gods), Sg. 39 b ; ace. tarsna deo (by the Gods), Sg. 217 b ; Fern. sing, dea, — in 
composition bandea (goddess), Sg. 60 a ; plur. bdndoe (goddesses), Sg. 53 b . 

II. Dia (day) : each dia (daily), Wb. 13 c ; indiu, hindiu (to day), Wb. ; fride, 
fridei (by day) ; dia brdiha (in the day of judgment), Wb. 23 c . 

III. Duine (man) — the radical ui becomes di in the plur. ; sing. gen. corp duini (a 
man's body), Wb. 12 a ; dat. donduini (to the man), Wb. 4 b ; ace. imfolngi induine 
sldn (he saves man), Wb. 4 d ; voc. a duini (0 man), Wb. l c ; plur. nom. inddini 

130 [xxxvi. Da sale is salivam tuam (da for du, do) ; im du da are, "around 
thy two temples" ; are (tempus capitis) gen. arach, is a c-stem. These examples 
are, therefore, improper.] 

131 [xxxvii. Rather do dib hguttib, where dib n=the Sansk. dwdbhydm, Greek 
dvolv (from dvo<piv).^\ 

180 Appendix. 

bi (the living men), Sg. 39 a ; gen. ice incheneli ddine (the salvation of the race of 
men), Wb. 26 d ; ace. corcefri dia et ddini (peace towards God and men), Wb. 20 d . 

IV. Ben (woman) — interchanges with the forms ban, mnd: iccfe inmndi (thou 
wilt heal the woman), Wb. 10 a . 

V. Rig (king) : sing. gen. itaig rig (in the king's house), "Wb. 23 b ; dat. ainm 
diarig (gl. Lar rex Vejentorum, i. e., the name of their king), Sg. 64 a ; plur. gen. 
hi lebraib rig (in the books of kings), Ml. 30 b ; ace. conroibtis ocdenum rectche 
la riga (gl. volentes esse legis doctores, i. e., to the kings),Wb. 28 a . 

VI. Za(day) is inflected from the forms Id, lae, and laithe, lathe (neuter). Sing. 
n. alaithe, Ml. 21 c ; gen. ammi maicc lai (we are the sons of day), "Wb. 25 c ; dat. 
illau bdiss (in the day of death), Wb. 29 c ; ace. fri laa brdtha (to doomsday), Wb. 
29 a ; plur. gen. ar lin laithe (in the number of days), Ml. 17 d . 

(B) Diminutives. 
Common to both, subst. and adj., like the declension of the first 
order. The instances that occur, especially in codex Sg., present the 
following terminations, -an, -en, and -that, which are more usual in 
the mas. and neut., and -ene, -ne, -not, -net in the fern. 

Masc. and neut. AN in substantives : duindn (a mannikin), Sg. 47 b ; tdiddn 
from tdid (a thief), 47 b . In adjectives beedn (gl. paullulus), Sg. 48 a ; trogdn 
(gl. misellus), 48 a . 

Numerous old proper names have the same ending: Tresan, Gibrian, Veran, 
Abran, Petran (vita S. Tresani, Boll., Febr. 2, 53). 

En : duinen (mannikin), Sg. 45 a . 

That: srdthathat (a sting), Sg. 47 a ; centat (gl. capitulum), 47 a . Chat, 
Nat, Net, are less frequent : duinenet (a mannikin), 45 b . 

Fern. Ene : lare'ne (from lair, a mare), Sg. 49 b . 

Nat in subst. : siurnat (gl. sororcula), Sg. 46 b ; talamnat (gl. terrula), 48 a . 

Net, Nit -.fochricnet (gl. mercedula), Sg. 47 a ; tonnait (gl. cuticula), 46 b . 

(Q) Degrees of Comparison. 

Comparative and superlative. The forms of the first, in the- old 
language, are the more copious, these are either regular or irregular, 


Of this there are two forms, -ithir, -iu, -u, — the first of which may 
be compared with the Creek orspog, and the second with the old 
Latin -ios, -ius, the s of which passed into r. Inflections are not 

Ithir I have only met in one codex Wb., and in one passage 27 d : islerithir. 

Iu and u are used indifferently, though the former is more usual in mono- 
syllables, the latter in polysyllables. The particle de is often met after the com- 
parative, corresponding seemingly to the Latin eo. 

Iu : nibia di mutaib bes huilliu inoensiU. (there cannot be more of mutes in one 
syllable), Sg. 7 a ; leriu (more industrious), 41 a ; semiu (more slender), 14 b ; goiriu 
(more pious), 40 b . 

U: oillu oldate cdiccet jer (more than fifty men), Wb, 13 d ; isassu, ba assu 
(easier), Wb. 15 c ; ata lobru (that are weaker), Wb. 12 b ; gliccu (wiser), Wb. 26 d ; 
istairismechu infer (the man is firmer), Wb. 28 b . There are some anomalous 
comparatives either in a, which sometimes becomes o, or with peculiar forms of 
their own. Of the former the principal are : — mda, mdo, mda, mdo (greater), 
messa (worse), nessa (nearer), tressa (stronger). Besides da (less), lia (more), 
ire (ulterior), ferr (better). 

Mda from adj. mar (great), for which mdr also occurs. From the form 
mar are produced mda, md, mdo : asmda alailiu (greater than another), Wb. 
12 a . From mdr are made mda, mdo, md : mda leu sercc atuile (greater with 

Zeuss on the Inflexions of Nouns. 181 

them is the love of their own will), Wb. 30 c ; Jresciu fogchricce asmoo (hope 
of the reward, which is greater), Wb. 10 c . 

Messa (worse) : fodaimid nech asmessa duib (endure one who is worse to you), 
Wb. 17 b ; creitmechsin asmessa ancreitmeck fthis believer is worse than an 
infidel), Wb. 28 d . 

Nessa (nearer) : isnesa do geintib (he is nearer to the Gentiles), Wb. 2 b ; innahi 
ata nessa (those which are nearer), Cr. 44 a . 

Tressa (stronger): combad tressa de hiress apstal do fulung (that the faith 
of the Apostles might be stronger to endure), Wb. 25 a ; ishe dim [din?'] ambe's 
adi inti diib bes tresa orcaid alaile (it is their habit that the stronger kill the 
weaker) Ml. 19 d . 

The three following comparatives, on account of the verbs accompanying 
them regularly in the sing., appear to have been originally substantives, with 
a comparative signification. They also sometimes act as adverbs in their naked 

Oa (less) : acoic indid oa q. xxx (the five in it less than thirty) Cr., 33 b . 

Lia (more, a greater number) : nabad lia diis no thriur dam (let there be not 
more than two or three) Wb. 13 a ; itlia sillaba o Ulitrib (there are more syllables 
of many letters), Sg. 71 a . 

Ire (ulterior): aither. ni ashire olddta m. ocus aui (patronymics no further than 
sons and grandchildren), Sg. 30 b . 

Ferr (better) : niferr neck alailiu and (no one better than another there), Wb. 
2 a ; nipatferr de (they are not better of it), Wb. 12 d . 

In the majority of the foregoing examples, the particle as, preceding the com- 
parative, is evidently the verb subst. 3 pers. sing, in dependent position. It is 
often, however, a different word, increasing the sense of the gradation, ex. gr. 
the comparative : ni asse acleith rafitir aslia (it cannot be easily concealed, many 
know it), Wb. 23 c , or of the superlative : asmaam. The meaning of compara- 
tive is still further increased by its repetition with the intervening formula ass : 
corrop moo assa moo et corrop ferr assa ferr donimdigidesseirc [donimdigid des- 
seirc] de et comnessim (so that it may be better and better, you increase your love 
of God and [your] neighbour), Wb. 23 b ; ferr asaferr (better and better), Wb. 15 c . 


There are two endings, -era and -am, the former of adjectives 
which form their comparatives in -iu -u, the latter of anomalous 
adjectives ending in a in the comparative. Internal inflexion occurs 
in the forms ending in am. 

Em : faillsem (most clear, lucid, front /o//ws, open, clear), Cr. 40 a ; tdisigem (the 
first ; in the verse : primus de Danaum magna comitante caterva), Sg. 42 a . 

The following are instances of the fuller form, -imem, -ibem, -bem after a double 
consonant or diphthong radical : huaislimem (the highest), Ml. 28 d ; itdoini saibi- 
bem dogniat inso (they are most false men who do this, — from saib false, or pro- 
perly delusive), M1.3 a . 

Am: oam (the least), Wb. 13 b ; asmaam ro'sechestar arsidetaid (it is he has 
reached as great an age as possible), Sg. 208 b , ata nessam (the nearest) Incant. 
Sg. comnesnam (the neighbour [lit. " nearest]), Ml. 36 a . Gen. : des&erc de et com- 
nessim (love of God and (our) neighbour), Wb. 23 b . Dat. : ho chomnesam (from 
a neighbour), Ml. 36 a . Ace. : galar bess fairechomnessam (the disease which is 
over on his neighbour). Cod. Camar. ; athis forachomnesam (reproach against his 
neighbour), Ml. 36 a . 



The following list of MSS. used by Zeuss in his Grammatica Celtica, and of 
the abbreviations he uses in referring to them, may be found useful to those 
who may not have that work : — 

Irish MSS. 

1. Codex Prisciani Sancti Galli (No. 904). A copy of Priscian preserved at 
St. Gall, and thickly interspersed with marginal and interlinear glosses. Zeuss 
denotes this MS. by Sg. 

2. Codex Paulinus Bibliothecae nunc Universitatis Wirziburgensis — marked 
M. th. f. 12. An MS. containing the epistles of St. Paul, formerly belonging to 
the cathedral church of Wurzburg, but now to the university. Zeuss refers this 
MS. to the eighth century ; although smaller than the St. Gall MS., it exhibits 
the same copiousness of glosses, if not greater, as they accompany the text after 
the manner of a continuous commentary, less by single words (like the St. Gall 
MS.) than explaining the context of St. Paul by Irish sentences. Zeuss denotes 
this MS. by Wb. 

3. Codex Mediolanensis bibliothecae Ambrosianae (marked C, 301, and 
denoted by Zeuss, Ml.), an MS. of St. Jerome's commentary on the Psalms, con- 
taining a mass of glosses, not less than the MSS. above mentioned. Zeuss 
agrees with Muratori and Peyron, that these commentaries were written by St. 
Columbanus, the founder of Bobbio, from whence the MS. was transferred in 1606. 
by Cardinal F. Borromeo, when he established the Ambrosian Library. Zeuss 
had only time to copy a small part of the glosses of this MS. 

4. Codex bibliothecae Carlisruhensis (No. 83, denoted by Zeuss, Cr.), an MS. 
formerly belonging to the monastery of Reichenau, containing " Computus de 
signis XII. et intervallorum ; Beda de ratione temporum". The text of Bede 
is interspersed with Irish glosses. 

5. Codex Prisciani bibliothecae Carlisruhensis, No. 223, denoted by Zeuss, 
Pr. Cr., a MS. of Priscian, also formerly belonging to Reichenau. It contains 
much fewer glosses than the St. Gall one, with which it agrees in part, and in 
part differs. 

6. Codex Sancti Galli (No. 1395), containing a collection of fragments from 
ancient MSS., made by Ildefonso von Arx, chief librarian. Folio 419, vol. II., 
has Irish formulae of incantation, hence Zeuss denotes it by Incant. Sg. 

7. Codex Civitatis Camaracensis (No. 619) an MS. belonging to the City 
of Cambray, containing the canons of an Irish Council held in 684 ; in one place 
in the middle of the book is preserved a fragment of an Irish sermon on self- 
denial intermixed with Latin sentences. The book was compiled for Alberic, 
Bishop of Cambray from 763 to about 790. 

Welsh MSS. 

1. The vellum MSS. in the Bodleian Library at Oxford,— Auct. F. 4-32, 
described by Wanley, in his catalogue of Anglo-Saxon MSS. 2, 63. The parts 
which supplied material to Zeuss being : a, a part of the grammar of Eutychius, 

Celtic MSS., etc. 183 

with interlined Welsh glosses, from p. 2 b to 9 a ; b, the exordium of Ovid's Art 
of Love, p. 37 a to 45 b , also containing interlined Welsh glosses. These glosses 
he believes to have been written in the eighth or ninth century, c, The alphabet 
of Nemnivus, giving the forms of the letters and their Welsh names ; d, some 
accounts of weights and measures in Welsh, intermixed with Latin, at p. 22 b 
to 23 a . 

2. A later vellum MSS. (Bodl. 572), containing theological tracts, and in the 
middle, from p. 4l b to 47 b , a list of Latin words with Welsh ones, which are 
written either over the Latin, or in the same line, with the sign .i. (id est), 
according to the custom of glossographers. 

3. The MS. of the Church of Lichfield (formerly of that of Llandaff), con- 
taining the Gospels, in various parts of which (e.g. p. 9 b 10b, 71 a , 109 b ), donations 
made to the Church of Llandaff at very ancient periods, not later than the 
glosses of the first Oxford MS., are noted in Latin, but with Welsh names, and 
even sentences, which Wanley has already published (p. 289). 

4. The Luxemburg Folio, a single leaf with Welsh glosses of the ninth cen- 
tury, which Mone found in the town library of Luxemburg, pasted to the 
cover of another MS. 

The glosses of these four MSS. have been published by Zeuss in the appendix 
to the Grammatica Celtica. 

5. Liber Landavensis, or book of Llandaff, compiled from more ancient docu- 
ments about the year 1 132. It contains many descriptions of boundaries of land, 
and also privileges of the bishopric, written in Welsh. It has also scattered 
through it Welsh proper names, especially of men and localities. This book Avas 
published in 1840 under the title, "The Liber Landavensis, Llyfr Teilo, or the 
Ancient Register of the Cathedral Church of Landaff", from MSS. in the 
Libraries of Hengwrt and of Jesus College, Oxford, with an English translation 
and explanatory notes by the Rev. W. J. Rees, published by the Welsh MSS. 
Society, Llandovery, 1810. 

6. Codex legion Veaedotianus, or the Venedotian MS. of laws. This MS., 
which belongs to the Hengwrt collection, is considered to have be en compiled in 
the twelfth century. The latest edition of these laws, the first collection of which 
is attributed to Hywel Dda (Howel the Good), who died a.d. 950, was pub- 
lished by the Record Commission, under the title " Ancient Law s and Institutes 
of Wales" ; comprising laws supposed to be enacted by Howel the Good, modified 
by subsequent regulations under the native princes prior to the conquest by 
Edward the First : and anomalous laws, consisting principally of institutions 
which by the statute of Ruddlan were admitted to continue in force. With an 
English translation of the Welsh text, 1841. 

7. The Red Book of Hergest (Llyfr coch o Hergest), now in the Library of 
Jesus College, Oxford. It is the chief of all the MSS., preserving the middle 
forms between the old and the living languages. Turner determined this MS. to 
be of the fourteenth century. The principal narratives relating to the history 
of Arthur and his Knights of the Round Table, which it contains, were published 
in three volumes, under the title of "The Mabinogion, from the Llyfr coch 
o Hergest, and other Welsh MSS., with an English translation and notes, by 
Lady Charlotte Guest". London, 1849. 

Cornish MSS. 

1. The vellum MS. marked Vesp. A. 14, in the Cotton collection in the British 
Museum. It is the most ancient monument which is known to exist of the 
Cornish language, and dates most probably from the twelfth century. It has 
been transcribed by Zeuss himself, and printed entire in the Grammatica 
Celtica, vol. II. 1100. It is also printed, more correctly, and arranged alpha- 
betically, by Mr. Edwin Norris, in the second volume of his Cornish Drama. 

2. A Cornish poem on the passion of Christ, of which four copies are extant, 
Of these, one is in the British Museum, and two are in the Bodleian Library. 
Both Cornish text, and an English version made in 1682, were published in 
London in 1826, under the title " Mount Calvary, or the history of the passion, 
death, and resurrection of our Lord and Saviour Jesus Christ", written in Cornish 


184 Appendix. 

(as it may be conjectured) some centuries past, interpreted into the English 
tongue in the year 1682, by John Keigwin, Gent., edited by Davies Gilbert. 
Mr. Whitley Stokes has recently published in the Transactions of the Philo- 
logical Society of London (1862) a new and corrected edition of this poem, with 
a translation, which is a^great boon, as the former edition was almost worthless. 

Armoric MSS. 

1. The Chartularies of the monasteries of Bhedon or Roton and Landevin. 
The former probably began at the end of the tenth or beginning of the eleventh 
century, and ended in 1 162 ; and the second in the beginning of the eleventh 
century. Those which have been printed will be found in Courson's Histoire des 
peuples Bretons dans la Gaule et dans les ties Britanniques, Paris 1846, and Dona 
Morice's Memoires pour servir de preuves a Vhistoire ecclesiastique et civile de 

2. The life of St. Nonna, or Nonita, a dramatic poem preserved in a paper 
MS., which was found by Marzinus, notary to the Bishop of Quimper, on 
Ids pastoral circuit, and presented by him to the editor. This MS., which 
Zeuss thinks belongs to the fourteenth century, has been published under the 
title : Buhez Santez Nonn, ou vie de Sainte Nonne, et de son fils Saint Bevy 
(David) Mystere compose en langue bretonne anterieurement au 12 me Siecle, 
publie d'apres un manuscrit unique, avec une introduction par VAbbe Sionnet 
et accompagne dhine traduction litterale de M. Legonidec et d^un facsimile du ma- 
nuscrit, Paris, 1837. 

Abbreviations used in Dr. EbeTs Celtic Studies. 

Irish words. All the Celtic words not specially distinguished by letters, 
whether quoted by Dr. Ebel or added to his lists, are Old Irish, and are taken 
from the Irish MSS. in the foregoing list. As the language of all of them is of 
about the same age — the eighth or beginning of the ninth century, — Ebel has 
not thought it necessary to indicate the particular MS. from which the word is 

Welsh words. "Words taken from the Welsh MSS. 1, 2, 3, and 4 in the fore- 
going list, are indicated by W. 1 ; those from 5 and 6, by W. 2 ; and those 
from 7, by W. 3. 

Cornish words. Words taken from the Cornish MS. 1 are indicated by V ; 
those from MS. 2, by P. 

Armoric words. All words taken from the Armoric MSS. 1 and 2 in the 
foregoing list, are indicated by the abbreviation Arm. 

The other abbreviations used by Dr. Ebel are : 

A.S.=Anglo-Saxon M. Ir.=Middle Irish 

Ch. Sl.=Church Slavonic N.H.G.=New or Modern High German 

Corn.=Cornish O.H.G.=01d High German 

Fr.=French O. Ir.=01d Irish 

Gaedh.^=Gaedhelic O.N.^Old Norse 

Gaul.=Gaulish O.S.=01d Saxon 

Goth.=Gothic Osc— Oscan 

Gr.=Greek Pruss.=Prussian 
Incant. Sg. refers to the Irish MS. No. 6 Sab.— Sabine 

K.=Kymric Slav, and Sl.=Slavonian 

Lat.=Latin Skr.=:Sanskrit 

Lett.=Lettish Umbr.=Umbrian 

Lith.=Lithuanian Z. refers to Zeuss' Grammatica Celtica ; 
Med. Lat.=Mediaeval Latin. the numbers to the pages. 

M.H.G.=Middle High German 

An * prefixed to a word indicates, as mentioned at Note 37, p. 60, that the 
word is hypothetical. The mark 2, used at p. 79 to indicate the degeneration of 
the case endings, is only an arbitrary sign. 



[The figures after the words, except where there is a special reference to a note, indicate 
the page; the italic letters after the figures refer to the columns, — a indicating the first or 
left hand column, and b the second or right hand column ; where these letters are not found, 
the word occurs in the general text. 

All the old Celtic and old Irish words which have been explained or analysed by Mr. Stokes 
in his Irish Glosses, are indicated by the letters St. ; the numbers which follow those letters 
without the letter p, refer to the numbers in his commentary on that work ; where the letter p 
precedes the figures, the latter refer to the page of that work. 

The letters C, I, w, after a word, indicate that it is also to be found in the list of Latin loan 
words given at p. xx. of Mr. Stokes' edition of Cormac's Glossary, published by Messrs. 
Williams & Norgate, London, 1S62. Whenever the word is spelled differently in Cormac from 
what it is in Zeuss, the word as spelled by the former is given in brackets.] 


akva, 113 
agni, 119 
antar, 1086 
ava, 1186 
avatya, 1186 
as (root), 112 
aina, 113 
ghaima, 112 
tauta, 113 
daiva, 113 
dacru, 113 
dant, 113 


agni, 119 

arm, 131 

antar, 1086 

anda, note 88, p. 112 

ava, 1186, 120 

as (i-oot), 112 

ayus, 109a 

avja, 113 

j an (root), 1106 

jna (root), 1106 

tamas, 113 

trna, 115a 

daia, 122 

dus-, 1086 


dydvvityog, 1166 
* ayaavixpog, 1166 
ayxh 109a 
ciyoj, 107a 

daru, 112 

diva, 113 

dus, 1086 

dha (root), 112, 128 

nava, 1116 

navy a, 1116 

nau, 1086 

panca (pancan), 122 

pra, 130 

bhu (root), 112, 128 

rad (?) (root), 1106 

vaskara, 112 


drc (root), 113a 
druh (roof), 1146 
dvara, 121 
dha (root), 112, 128 
nava, 1116 
navya, 1116 
nau, 1086 

panca (pancan), 122 
pibami ( Ved.) 1086 
pra, 130 

bodhayami, 117a 
bhu (root), 112, 128 
mahat, 121 
lax (root), 118a 
lap (root), 1186 
vata, 1146 

vid (root), 112 
vidhava, 113 
visha, 1086 
vira, 113 
sak (root), 113 
sana, 113 
sarpa, 109a 
sauala [?], 113 
su-, 1096 
sru (root), 112 
svastar, 113 

vid (root), 112 

vidhava, 113 

visha, 1086 

vira, 113 

vrhat, note 82, p. 97 

v6da, 123 

cru (root), 112 

sarpa, 109a 

su-, 1096 

sru (root), 112 

srotas, 1116 

Old Persian. 

vazarka, note 82, p 97. 



cieXka, 1076 

drjp, 107a 

amy, 109a 

afekiOQ, v. i}\iog, 112a 

aKovaag, 129 

dXeoj, Ilia 

aXkdaao), 107a 
aWofiai, 109a 
dXXog, 109a 
aXoyia, 124 
a\oyo£, 124 
&\g, 1116 



Indices Verborum to Position of the Celtic. 

djifyi, 99a 
avaTrndrjffag, 129 

dvCLGTCtQ, 129 

dvrjp, 1086 
avoTrXog, 124 
dvrl, 99a, 123 
aoirXog, 124 
d-rrofiaXuv, 129 
arroKpidug, 129 
areicvog, 124 
-arog, 126 

avxfJ-og, v. aavx^og, 112a 
avxuvpog, v. tjavxfiog, 

/3dva (Boeot), 1096 
/3io£, 1096 
/3ou£, 108a 
j3paxi<i>v, 108a 
/3p£%w, 117a 

ysv£0\ov, 1106 
ytwaw, 1106 
yEpavog, 113a 
y'iyvo\iai, 1106 
ytyvuxTKU), 1106 
ypdcjxo, 1026, 115a 
ypaug, 1086 
yuvatKOf, 124 
yyj/?7, 1096 

Sdicpv, 110a 
dafid^oj, 110a 
dedefi&vov, 129 
fo£i6e, 110a 
depKU), 113a 
flop?, 1086 
#6pt>, 110a 
flpug, 110a 
fog-, 1086 

eyyvg, 109a 
U, 1076 
?0oe, 1136 
fl^ap, 1106 
e7/ii } 129 
hvaXog, 1086 
eicdg , 109a 
t/cypog, c/cupd, 1116 
ZXafiov, 129 
eXatrtrwj/, Ilia 
sX/iivg, 110a 
evn, 112a 
ivvstte, 112a 
Eirojiai, 112a 
Ipiov, 1126 
ipn-frov, 109a 
-£ff, 124 

££7T£pof, 1126 
eralpog, 112a 
sVapog, 112a 
tv, 1096 

ttp%, 1086 
?wyo£, Ilia 

i]j3r}, Ilia 
riQog, 1136 
??i0£O£, 1126 
ijXiog, 112a 

0ai/ (roo*)> 122 
Osog, 110a 
^rog, 122 
Ovpa, 1106, 121 

t£w, 112a 
log, 1086 

'iTTTTOg, 1106 

-((Taa, 124 

Ka(3dXXr]g, 110a 
Kd/jnrru), 110a 
Kapdia, 1096 
KarEpyaadfiEVOV, 129 
kevOoj, 110a 
KvrjfXT], 108a 
Kvu)v, 110a 

\dj3w, 129 
Xay%dvw, 113a 

\a'iy£> 1086 
\afi(3dv(o, 129 
\fi%w, 1116 
Xi0oc, 1086 
\6yoc, 124 
Xoyx*], 1086 
Xvaavreg, 129 

fikyag, 121 
/tedofxai, 1086 
/i£\e, 1086 
/f£ig, (Ion.), Ilia 
fikfiova, Ilia 
/xt]Xivog, 1196 
fiiayu), 1116 
/u»7v, v. )UB£, Ilia 
fxrjrrjp, Ilia 
IxvXr], Ilia 

va£f, 1086 
v'eeiv, 1086 
VE<pkXr}, 1116 
i^0«v, 1086 

6£ot>c, HOa 
6dvvr)<parog, 1136 

o~licog, 1126 
bvofxa, 107a 
o/cwf, 231 

6p0o C (Fop0og),lO7& 
-og, 124 

rraXafir), 1116 
7rarrjp, 1096 
TrsAfKWf, 1046 
7repidepicr)g, 113a 
7rE([>r)<T0fiai, 1136 

7TS(pVOV, 1136 

7r\drof, Ilia 
irXarvg, Ilia 
7rXr]^jxvpig, 1116 
7rvp, 119 

Tcvpyog (tyovpicog), 1136 
7rwXov, i29 

p£W ((TpfiFw), 1116 

aavKog, 1116 

<ravxpog,v. avxfJ-og, 112a 
aavcrap6g,v.aavKog, 112a 
cncorof, 1136 
(rpsFoj, v. pah), 1116 
(rra'e, 129 

<srpv\iov ( Thracian), 1116 
(TvvaxOsvrojv, 129 
o<pvpov, 1086 

raupog, 1096 
ravaof, 1126 
raw-, 1126 
-rarog, 126 
7-£yog, 1126 
repErpor, 1096 
Tspxvog, 1156 
rerra, 1096 
-rog, 126 
rp£%w, 113a 
rpvx^i 1136 

vfxihv ,129 
i)7H/0£, 1116 

(paXXog, 1076 
0ei/, 1136 
<piXog, 116a 
<p6vog, 1136 
ippqrrip, 1096 

Xaivo), 1136 
XEifjaov, 1106 
%twv, 1106 

wXevj?, 1106 
yoi>, 1126 

Latin Index, 




abbas, 99a 
abecedarium, 99a 
abstinentia, 99a 
accentus, 99a 
accidens, 99a 
accoinmodatio, 106a 
aceo, 99a 
acer, 99a 
acetum, 99a 
actualis, 99a 
acuenda esset, 99a 
acutus, 99a 
ad-, 113a 
adigit, 107a 
adjectivum, 99a 
ador, 1086 
adorare, 99a 
adoratio, 99a 
adulterium, 99a 
adversarius, 99a 
aer, 107a 
aetas, 1086 
aevum, lc9a 
agnus, 1096, 1116 
ago, 107a 
alius, 109a 
ala, 113a 
alo, 1 136 
altare, 99a 
altura, 99a 
amnis, 1076 
anacboreta, 99a 
ancora, 99a 
angelus, 99a, 114 
angor, 109a 
angustia, 109a 
angustus, lU9a 
anima, 107a 
animal, 99a 
apostolus, 99a, 114 
appetere, 1046 
applicare, 99a 
arare, 1096 
aratrum, 99a 
arcbiepiscopus, 102a 
arduus, 1076 
argentum, 99a 
-arius, 124 
arma, 996 
arrnilla, 996 
ars, 996 
articulus, 996 
artus, 107a 
arvum, 107a 
asinus, 996 

atomum, 996 

auctoritas, 996 

augusti, 99a 

aura, 107a 

aurura, 996 

avignus, v. agnus, 1096 

axilla, 113a 

baculum, 996 
badius, 108a 
balbus, 996 
baptista, 996 
baptizo, 996 
baptisnia, 996 
barba, 996 
barca, 996 
basilica, 100a 
basium, 100a 
battuere, 100a 
beat us, 100a 
benedico, 100a 
benedictio, 100a 
benedictus, 100a 
bestia, 100a 
betula, betulla, 10Sa 
blasphemare, 100a ' 
bibo, 1086 
bos, 108a 

bracbium, 100a, 108a 
brassica, 100a 
brevis, 100a 
broccus, broccbus, 100a 
bulla, 100a 
buxus, 100a 

caballus, 110a 
cadere, 108a 
caecus, 1136 
calamus, 100a 
callidus, 100a 
calix, 100a 
camisia, ICOa 
cancella, 100a 
cancellarius, 100a 
cancer, 100a 
candela, 100a 
candelarius, 100a 
candelabrum, 100a 
cano, 108a 
candidus, 108a 
canis, 110a 
canon, 100a 
capellanus, 100a 
caper, 114a, 123 
capere, 123 
capio, 114a 
capistrum, 100a 

capitulus, 100a 
capra, 114a, 123 
captus, 100a 
caput, 100a 
carbunculus, 100a 
career, 100a 
caritas, 100a 
car(o)enum, car(o)enaria, 

carpentum, 1006 
carus, 108a 
caseus, 1006 
castellum, 1006 
castra, 103a 
castrum (for cad-truni v. 

note 85), 10Sa 
castus, 1006 
castitas, 1006 
catena, 1006 
cathedra, 1006 
catholicus, 1006 
caucus, 1006 
caules, 1006 
causa, 1006 
cedere, 100a, 108a 
cedria, 1006 
cella, 1006 
celo, 1146 
census, 1006 
cera, 1006 
cervus, 108a 
cervical, 1006 
cervisia, 1006 
character, 1006 
chorda, 1006 
christianus, 1006 
chrisma, 1006 
cilicium, 1006 
circare, 1006 
circinus, 1006 
circulus, 1006 
circumflexus, 1006 
ci vitas, 1006 
clarus, 1006 
classis, 1006 
claustrum, 1006 
clavi, 108a 
clericus, 1006 
clima, 1006 
coccus, 1006 
coloni, 1006 
columba, 1006 
columella, 1006 
columna, 1006 
*cominitiare, 101a 
commatres, 101a 
commixtio, 1116 

188 1 .: J ictS Verborwm to Pt \ - ■ r the Ce 

commodum. K la 
communio, 101a 
compar. 101c 
com-; uativus, Ilia 
compatres. 1 ". 1 
concedere 1 l« 
confessio, 101a 
confligere, 101a 
confortare, 101a 
consecratio, 101a 
consilium, 101a 
consimilis. 10J 
consona, 101a 
conventus. 101 
coquina, 101a 
coquus, 101a 
cor, 1 

corona, 101a 
eoronatus. 1 . 1 
corpus. 1 - 
corrigia, 101 
corylus. 1. 1 
coryletum, 101a 
costa, 10 r 
coxa v. costa. 1 
craticula, 101a 

3r, 101a 
creatura, 101a 
credo, 1 
credulus, 101a 
crepusculum, 101a 

. la 
cribruin, 101 - 

erudelis, 101a 



cuprum, 1 
cuttus, 1 1 1 
; , 110a 
cymbalum, 1016 
as, 1016 

..:±a, v. lacrima, 110 
daem i 
damnare, 1016 


decedere, 10 1 1 
decima, 1 
defendere, M 
denarius. ! 

dens. 110a 

deprecatio, 1016 

deseendere, 11 I 

liesidexabat, 10W 

]fsx . ■::':.?. 11 1: 

deus, 11( 

dexter, 110a 

diaoolus. 1016 

diaconus, 1016 

diaconissae, pi, K 

dictator, 1016 

dies, 110a 

diesjovis, 111 

lies soils, iOU 

digamma, 1016 

digitus, 1011 

diluvium. 1 

dinsnia, v. linsua. 11. . 

discere, 1 '. 
discipulus. 1016 
discretua, 1116 
discus. 1 : 

doctus, 101 
dolor, 1016 
dominica. I 1 - 
donio. 11 1 
draco, 1016 
drungus, 115a 

dubitare. 1014 

dubi d 

dubitantia, 1016 
durus, 1016 

ecclesia. 1 I 

i : 

eleera osyna. 1016 
emendare, 103 
-ensis. 124 
episcopus, 102 a 
epistola, 102 a 
equus, 1106, 122 
eremita, 102a 
esculus, 102 
esox, 102a 
etymologia, Il- 
ex. 107ft 

excomrnunicatus, 102a 
evangelium, 102a 

faba, 1 - 
facies, 102 
iagus, 102 
faginus, 102 

falco, I 

: - 

favere, 102 

i;ailraaa:n. I 1 ' . 
fen-srra, 102 
ferre. 109J 
fibula, 102a 
ficus. 102 
fieulnus. 102a 
fides, 102 i 
figura, 102 a 
finis, 102 a 
±L-^:raie:::ura. Ml 
flagelium, 1 
fianrma. 102 a 
flecto. 102a 
flos, 11". 
foeniculum. 1 
fores. 1106,121 
forma, 102 a 
fossa. 102* 
frazrare. 1021 
frater, 1096 
frenmn. 102 
fructus. H .' 
fbgere, 102 
i 102* 
fulgur, 102 
funis. 102 
fur, 1026 
furca, 103 
furnus, 102 
fastis, 1026 

geminantur. 102 
gentes, li- 
as, 1026 
gentilitatis. 1 _ 
genitivus, 102 
gens. 1 : . 
gigno, 1106 

gradale. 102 
gradus, 102 
gratia (sratias agimas), 

1026 " 
grarari, 1026 
gravis, 1026 
grus, 1 

habilis, 102 
haeretici, pi 102 
hastula. 102 
hiare. 11! 
hiems, 11 
bistoria. 1 - 
honor, 102 
bora. I 
hosp; - 102 

Latin I tided'. 

18 l J 

humilis. 1026 
humilitas, 1< \ 
Iniunus, 1026 

idolum, 1026 
idus, 1026 
ignis, 119 
imago, 1026 
iinpedicare, 1026 
irnperator, 1026 
itnperium, 1026 
improbitas, 105a 
improbus, 105a 
incensum, 103 
indupedio. note SO, p. 120 
iaduperator. note 89, p. 120 
inermis (*ineraiiusj, 12-1 
infamis, 103a 
infernuni, 103a 
infinitivus. 103" 
initium. II 
insece, 112a 
instrumentum, 103a 
insula, 10S6 
inter, 101 
interjectio, 103a 
inter rare, 103a 
-issimus. 126 
-isti, 128 
-istis, 12S 

jejuniuni, 103a 
judex. 103^ 
juguni, Ilia 
jusculum, 103" 
justitia, 103a 
j uveitis, Ilia 
juvencus, Ilia 

kalendar, 103a 

labes, 1086 

labo, 1086 

labor, 1086 

lac, 10S6 

lacrima, v. dacrima, 110a 

laicus. 103a 

lana, 1126 

lancea, 1086 

lapis, 10S6 

latex, 103a 

latro, 103a 

latus, Ilia 

laurus, v. daurus, 110a 

lector, 103a 

lectus, 103a 

legalitas, 103a 

legere, 103 a 

legio, 103 a 

leo. 103a 

levis, Ilia 

liber, 103a 

ligo, 103a 

lilium, 103a 

lingo, Ilia 

lingua, v. dingua, 110a. 

1146, 1086 
linucu, 103" 
liquida, 103" 
littera, 103a 
liveo, 114" 
lividus, 1 1-4" 
livur. 114a 
loculus, 103a 
locus, 103a 
locusta, 103a 
longa, 103a 
longa (navis), 103a 
lorica, 10! 
lucerna, 1 
lunaris, 103a 
lutum, K 

magister, 103a 
ma gnus, 121 
«r, 103" 
maleJictiu. 1 
maledictus, 103a. 
rnalitia. I 1 
malva. 103" 
mancus, 1036 
manere, lo3 / > 
manna. 103^ 
mantellum, 103- 
manus, 1036 
mare, 1116 
margaritae, 1036 
martulus (martellus),1036 
martyriuni, l( 
masculinum, 1035- 
mater. Ilia 
matutinus, 10S6 
medicus, 1036 
medicina, 1036 
meditor, 1086 
medius, medium, 111a 
mel, 1086 
membra, 1036 
memini, Ilia 
memoria, 1036, Ilia 
mendicus, 1036 
mensa, 1036 
mensis, 1116 
mensura, 1036 
meretrix, 1036 
metrum, 1036 
mille, 1036 
millefolium, 1336 

miles, 1036 
militia, 1036 
ministrare, 1035 
minus (minus facere), 1 
mirabile, 1036 
miraculum, 1036 
mirus, 103^ 
misceo, 1116 
modus, 1036 
molina, 1036 
molo, Ilia 
monachus, 1036 
monasterium, 1036 
mons, 1086 
moralis, 1036 
mori, Ilia 
morticinium, 1036 
mortuus, Ilia 
morus, 1036 
mulus, 1036 
mulxi. Ilia 
murus, 1036 
muta, 1036 
myrias, 1036 
mvrtus, 1036 

nascor (gnascor), 1 106 
natalicia, 1036 
nates, 104a 
natio, 104a 
natrix, 114a 
navis, lu3a, 10*6 
nebula, 1116 
negotium, 104 
nepos, 114a 
neptis, 114a 
nere, 1086 
nerio, Xero, 1086 
neutrum, 104a 
nidus, 114a 
*nigvis, nihvis. v. nix. I 
rimbus, 104a 
nix, nivis, 1166 
nosco (gnosco), 1106 
nota, 104a 
notarius, 104a 
novellus, 104a 
norus, 1116 
nox, 1116 
nudus, 116a 
numerus, 104a 
nuptiae, 104 a 
mix (cnux), 110a 

obediens, 104a 
oblatio, 104a 
occulo. 1146 
octo, 122 

oenos, O. L., 1116 
offeree, 104a 


Indices Verborum to Position of the Celtic. 

olea, 104a 
olor, 107a 
operarius,- 104a 
optativus, 104a 
opus, 104a 
oracuhim, 104a 
orate, 104a 
oratio, 104a 
ordinare, 104a 
ordinatio, 104a 
ordino, 104a 
ordo, 104a 
ostiarius, 104a 
ostreurn. 104a 
ovum, 1126 

paganus, 104a 
pagus, 104a 
pallium, 104a, 1146 
palma, 104a, 1116 
palus, 104a 
panis, 104a 
papa, 104a 
papilio, 104a 
papyrus, 10 la 
paradisus, 104a 
parare, 104a 
paries, 104a 
parochia, 104a 
pars, 104a 
pascha, 104a 
passio, 1046 
patella, 1046 
pater, 101)6 
paucus, 1186 
pauper, 1046 
pausa, 1046 
pavo, 1046 
pax, 1046 
peccatum, 1046 
pedester, 1046 
pelliceus, 1046 
pensus, 1046 
pentecoste, 1046 
penultima, 1046 
peregrinus, 1046 
perfectus, 1046 
pergaminum, 1046 
persona, 1046 
petere, 1046 
phiala, 1046 
philosophus, 1046 
philosophia, 1046 
piimaculum, 1046 
pinus, 1046 
piper, 1046 
pirus, 1046 
piscis, 1046, 114a 
piscator, 1046 
pistor, 101a 

pistrinum, 101a 
plaga, 1046 
plangere, 1046 
plenus, 1046, Ilia 
plebs, 1046 
plicare, 1046 
plum a, 1046 
poena, 1046 
poenitere, 1046 
poenitentia, 1046 
pondo, 105a 
pons, 105 a 
populus, 105a 
porcellus, 105a 
porcus, 105a 
porta, portus, 105a 
portare, 105a 
positivus, 105a 
postilena, 105a 
postis, 105a 
praebendarius, 105a 
praeceptum, 105a 
praedico, 105a 
praelatus, 105a 
praeservare, 105a 
prae stare, 105 a 
prandium, 105a 
presbyter, 105a 
pretiare, 105a 
primus, 105a 
princeps, 105a 
prior, 105a 
probabitur, 105 a 
probatus, 105a 
probus, 105a 
prologus, 105a 
pronomen, 105a 
propositus, 105a 
propheta, 105a 
proprius, 116a 
prudens, 105a 
psalmus, 105a 
psalterium, 105a 
purgatorium, 105a 
purpura, 105 a 
purus, 105a 
puteus, 105a 

quadragesima, 105a 
quaerere, 1086 
quaestio, 105a 
quinquagesima, 105a 
quiritare, 1026 

rastrum, 105a 
rectus, 1146 
reddere, 1056 
regnare, 1056 
regula, 1056 
reliquiae, 1056 

remus, 1056 
rete, 1056 
rex, 1056, 114a 
rigo, 117a 
rogavi, 122 
rosa, 1056 
rosetum, 1056 
rota, 1116 
ruber, 1116 
ruta, 1056 

sabbatum. 1056 
saccus, 1056 
sacerdos, 1056 
sacrificium, 1056 
sacrilegium, 1056 
saeculum, 1056 
sagita, 1086 
sagum, 109a 
sal, 1116 

salicastrum, 1056 
salio, 109a 
saliva, 1056 
salix, 109a 
saltus, 1056 
salutare, 1056 
salvare, 1056 
sanctus, 1056 
scabellum, 1056 
scala, 1056 
scandere, 1056 
schola, 1056 
scbolasticus, 1056 
sciens, 1056, 
scribere, 1056 
scrinium, 1056 
scripulus, 1056 
scutella, 1056 
scutum, 109 a 
sebum, 1056 
securus, 1056 
secus, 109a 
sedeo, 112a 
senator, 1056 
senex, 112a, 124 
senior, 1056 
sensus, 1056 
sepelire, 1056 
sepultura, 1056 
septimana, 1056 
septuaginta, 106a 
sequor, 112a 
sermonarius, 106a 
serpens, 109a 
serus, 106a 
sextarius, 106a 
siccus, 1116 
signum, 106a 
similis, 109a 
situla, 106a 

Latin Index. 


socer, 1116 
socrus, 1116 
sol, 112a 
solarium, 106a 
solitarius, 106a 
somniari, 106a 
somnus, 1116 
sophist a, 106 a 
soror, 112a 
sors, 106a 
-sos, 127 
spatium, 106a 
sperare, 106a 
spina, 106a 
spiraculum, 106a 
spiritus, 106a 
spoliare, 106a 
spongia, 106a 
sponsa, 106a 
stabulum, 106a 
stagnum, 106a 
stannum, 106a 
status, 106a 
stimulus, 106a 
stlocus (O. L.), 103a 
stola, 106a 
stragulum, 106a 
strata, 106a 
strigilis, 106a 
-sum, 127 
superlativus, 106a 
sus, 118a 
syllaba, 106a 
synodus, 106a 

tabellarius, 106a 
taberna, 106a 
tabes, 106a 
tacere, 1146 
talentum, 106a 
tardare, 106a 
-tas, 124 
taurus, 1096 
tellus, 106a 
temere, 1126 
tempero, 1066 
templum, 1066 
temptare, 1066 
tendere, 1066 
tenebrae, 1126 
tenuis, 1126 
terebra, 1096 
terminus, 1066 
terra, 1096 
tertia, 1066 
testis, 1066 
testimonium, 1066 
theca, 1066 
theoria, 1066 
thesis, 1066 

thronus, 1066 

thus, 1066 

Titan, 1066 

titulus, 1066 

torques, 1066 

torrens, 1116 
j torta, 1066 
I totus, 1066 

tractus, 1066 

traditio, 1066 

trans, 1096 

tribunus, 1066 

trinitas, 1066 

fcripus, 1066 

tristis, 1066 

tructa, 1066 

truncus, 1066, 1156 

trux, 1146 

tuba, 1066 

tugurium, 1126 

tunica, 1066 

turba, 1066 

turris, 1066 

tympanum, 1066 

ulna, 1106 
ultima, 1066 
uncia, 1066 
unctare, 1066 
unguere, 1066 
unicornis, 1066 
unus, 1116 
ursus, 1066 

vacca, 118a 
vagina, lu66 
vates, 1096 
velum, 1066 
venenum, 1066 
ventus, 1146 
vermis, 110a 
versatile, 1066 
veru, 1076 
versus, 1066 
verus, 1066 
vesper, 1126 
vetus, 107a 
vicus, 1126 
vidua, 107a, 1126 
vigil, 107a 
villani, 107a 
vinea, 107a 
vinum, 107a 
viperae, 1076 
vir, 1 126 
viridis, 1096 
virus, 1086 
virtus, 1076 
visio, 1076 
vita, 1096 

vitium, 1076 
vivus, 1096 
vocula, 1076 

Mediaeval Latin. 

baro, 100a 
brace, 100a 

caldaria, 100a 
cattus, 1006 
clocca, 1006 
cloccarium, 1006 
companium, 101a 
conucula, 101a 

follis, 102a 
fontana, 102a 
foresta, 102a 
forestis, 102a 

gridare, 1026 

hanapus, 1026 

melinus, 1196 
mirare, 1036 
multo, 1036 

padulis, 104a 

sappetus, 1056 
sicera, 106a 
solta, 106a 
stratura, 106a 

torneamentum, 1066 


Auximum, 118a 


ausum, 996 
nerio, 1086 


ner, 1086 
nesimo, 114a 
teerum, 1096 
tuvtu, 1126 


berus, 1076 
berva, 1076 
esme, 125 
esmei, 125 
nesimo, 114a 

192 Indices Verbornm to Position of the Celtic, 

ner, 1086 
pir, 119 
pusme, 125 
toto, 1125 
traf, 1096 


-esis, 121 
finisco, 102a 
^issa, 121 
peso, 1046 
putana, 105a 
rendere, 1056 
stendardo, 106a 


landa, 118a 
veltro, 107, 118a 



landa, 118a 


bai, 108a 
blamer, 100a 
broche, 100a 

cbarite, 100a 
cherir, 108a 
commencer 101a 

empecber, 1026 
enceiis, 103a 
estaminet, 106a 
estonner, etonuer, 102a 

foudre, 1026 

baster, bater, 1026 

lande, 118a 

merveille, 1036 
mesfaire, rnefaire, 1036 
moi, 126 

pan, 123 
parfait, 1046 
paroi, 104a 
pavilion, 104a 
petit, 1046 
pommaille, 105a 
prison, 105a 

rame, 1056 
recommendare, 1056 

sauver, 1056 
sou, 106a 
songer, 106« 
sorte, 106a 
soutenir, 106a 

talent, 106a 
tonneau, 1066 
trabison, 1066 
tribunal, 1066 

vice, 1076 


ada (Kr. Goth.) note 88, 

page 112 
* ADDIA (Prim.), note 

88, page 112 
afvairpands, 129 
aggilus, 114 
ains, 1116 
aivs, 109a 
alan, 1136 
aleina, 1106 
alis, 109a 
alja-, 109a 
aljan, 1136 
alls, 1186 
anabiudan, 117a 
and-, 123 

andbindandans, 129 
anbafjands, 129 
apaustaulus, 114 
ara, 1156 
arbi, 1166 
arbja, 1166 
-areis, 124 
at-, 113a 
aubsans, 118a 
auburoa, 118a 
aubumists, 118a 
auf>eis, 1186 

balgs, 117a 


ban] a, 1136 
baurd, 117a 
baurgs, 1136, 117a 
braids, Ilia 
broj^ar, 1096 

dailjan, 115a 
dails, 115a 
daur, 1106, 121 
dauro, 1106 
driugan, 1156 
du, 1156 

eisarn, 118a 

fadar, 1096 

faurbiudan, 117a 

fimf, 122 

fisks, 114a 

fotubaurd, 117a 

fram, 126 

frunia, 126 

fulan, 129 

fulls (i.e. fubis), Ilia 

fulljan, Ilia 

ga-, 130 
gaarbja, 1166 
gabundanana, 129 
gadraubts, 1156 
gagaggandam, 129 
gabausjands, 129 
gastandands, 129 

gataujandan, 129 
gazds, 1176 
gild, 1176 
graban, 115a 

bafja, 114a 
haihs, 1136 
bairto, 1096 
haurn, 114a 
bilpan, 123 
buljan, 1146 
bunds, 110a, 123 

iddja, 128 
idreiga, 116a 
is, 126 
ita, 126 
-iza, 126 
Tzvis, 129 

jains, 127 
juggs, Ilia 
jubiza, Ilia 
juk, Ilia 

kan, 1106 
kuni, 1106 

laigo, Ilia 
lamb, 118a 
land, 118a 
leihts, Ilia 

Old High German Index. 


mag, 1156 
manags, 116a 
managei, 116a 
man, Ilia 
mena, 1116 
menojps, 1116 
mes, 1036 
mikils, 121 
missa, 118a 

nadr, 114a 
namo, 109a 
naqvaj^s, 116a 
nehv, 114a 

og, 116a 
ogan, 116a 
-oza, 126 

qvens, 1096 
qvino, 1096 
qvius, 1096 

raihts, 1146 
reiks, 114a 
rign, 117a 
runa, 1186 

saian, 1166 

salt, 1116 

salta, 109a 

sakan, 112a 

sarnajj, 129 

sauil, 112a 

si, 126 

sidus, 1136 

sineigs, 112a 

sinista, 112a 

sinj^s, 1186 

sitan, 112a 

skadus, 1136 

skalks, 1186 

* snaigas, * snaigvas 

snaivs, 1166 
snaivs, 1166 
sokjan, 112a 
stiur, 1096 
surma, 1186 
sunno, 1186 
sumis, 121 
svaihra, 1116 
svaihro, 1116 
svistar, 108a, 112a 

tagr, 110a 
taihsvs, 110a 
tarn j an, 110a 
timan, 110a 
triu, 110a 
tuggo, 1146, 123 

tunjms, 110a 

J?ahan, 1146 
)?airh, 1186 
f^aurnus, 1156 
J?iuda, 1126 
jpragja, 113a 

ushlaupands, 129 
usstandands, 129 

vair, 1126 
vairj?s, 1166 
vait, 123 
valdan, 1166 
vaurms, 110a 
veins, vehs, 1126 
viduvo, 1126 
vityeis, 1196 
vinds, 1146 
vulfs, 121 
vulla, 1126 

Old Teutonic. 

* hafar, 123 
hafjan, 123 

Old High German. 

ahsala, 113a 
ali-, 109a 
angi, 109a 
angil, v. engil, 114 
aphul, aphol, 115a 

baga, 117a 

bagan, v. biag, 117a 

bagen, 117a 

bana )t /, 1136 

bano, m., 1136 

banon, 1136 

bart, 996 

biag, v. bagan, 117a 

bigil, v. bihal, pigil, 117a 

bihal, v. pihal, 117a 

bimunigon, 124 

biseh, 1116 

bisihan, 1116 

blat, v. plat, Ilia 

bluot, 117a 

boch, v. poch, 117a 

bogo, v. poco, 117a 

bort, 117a 

borti, 117a 

borto, 117a 

charra, v. karra, garra, 

chona, 1096 

chraft, 1176 
chrannb, 113a 
cbrump, 1176 
chruzigon, 124 
chuo, 1096 
cbus, 1176 

dach, 1126 
demar, 1126 
dunni, 1126 

ecala, 1176 

egala, 1176 

ei, 1126, note 88,/?. 112 

elitliiotic, 109a 

engil, v. angil, 114 

ewa, 109a 

fiur, 119 
folma, 1116 
friudil, fridil, 11 6« 

gabala, 1176 

galingan, 113a 

ganzo, 1156 

garra, v. karra, 1176 

gart, 1176 

gartja, 1176 

ger, 1176 

ginen, 1136 

ginon, v. ginen, 1136 

gisal, 1176 

giwiznes (neut.), 119a 

giwiznesi (Jem.), 1186 

glas, 1176 

grioz, 1176 

hadu, 1176 
helan, 1146 
heli, 1146 
hiruz, 108 a 
hnot, 110a 
bosa, 1176 
nulla, 1146 
hut, 110a 
hutta, 110a 

isarn, 118a 
iwa, 118a 

karra, v. garra,chirra, 1176 
kramph, 1176 
kruog, 115a 
krus, 1136 

ledar, 118a 
luogen, 118a 
lus, 118a 

mana (ma.nh.a-), 111 
maracb, 118a 


Indices Verborum to Position of the Celtic. 

meriha,/. v. marach, 
mias, 1036 
niiscjan, 1116 
muotar, Ilia 

nachat, 116a 
natra, natara, 114a 
nefo, 114a 
nest, 114a 
nibul, 1116 
nift, 114a 
niftila, 114a 

pale, 117a 
phant, 123 
pigil, v. bigil, 117a 
pihal, v. bihal, 117a 
plat, v. blat, Ilia 
pli, 114a 
pliwes, 114a 
poch, v. boch, 117a 
poco, v. bogo, 117a 
postul, 114 

rad, 1116 
reht ? 1146 
richi, 114a 
run, 1186 

sagen, 112a 
sech, 1186 
segal, 1186 
sia, 127 
sie, 127 
sind, 1186 
siniscalc, 112a 
sio, 127 
siu, 127 
snecco, 1186 
snuor, 1186 
stroum, 1116 
sumar, 1186 
sumna, 1186 
sunna, 1186 

tarch, 1176 
tiligon, 124 
triugan, 1146 
truhtin, 1156 
trubtinc, 1156 

wagan, 1186 
wer, 1126 
weralt, 1126 
witu, 119a 
wolchan, 1196 
wolf, 121 

za, zi, zuo, 1156 
zand, zan, 1 LOa 


zi, v. za, 1156 

zoraht, v. zhort, 113a 

zorht, 113a 

zorft, v. zohrt, 113a 

zun, 1176 

zuo, v. za, 1156 

Middle High German. 

bil, 117a 

bluot,/)/. bliiete, 117a 
bader, 1176 
limpfen, 118a 
man, 118a 
march (marc), 118a 
vluor, 118a 
vriedel, 116a 

New High German. 

aufgebot, 117a 
ausgesprocben, 130 

bemachtige, 124 
blappen, blappern, v. plap- 
pern, 1186 

enge, 109a 
enterben, 117a 

gefahrtin, 1186 
genter, 1156 
gerte, 1176 

hader, 1176 
haksch, 118a 
hulle, 1146 

jemand, 123 

kind, 1106 
krug, 115a 

lahm, 118a 
lugen, 118a 

macbtig, 124 
maultbier, 100a 
menge, 116a 
mis-, 118a 
mucke, 118a 
muscbel, 114 

ode, 1186 

peinige, 124 

plappern,u. blappern, 1186 

reinige, 124 

walten, 1166 

werth, 116a 
windbeutel, 102a 

zahlen, 115 

Old Saxon. 

ehu, 1106 
reht, 1146 
tun, 1176 
vidu, 119a 
wolcan, 1196 


appel, 115a 

Low German. 

kaute, kute, 105a 

Middle Hutch. 

slecke, 1186 


agg, note 88,/>. 112 
appel, 115a 

bat, 117a 
boga, 117a 
bucca, 117a 

coss, 1176 
craft, 1176 
crumb, 1176 

deorc, 1176 

eorl, 115 

flor, 118a 
folma, 1116 

gandra, 1156 
gar, 1176 
gevitnesse, 119a 
gevitnes, 119a 
gia's, 1176 
greot, 1176 

hafer, 114a 
heaSo, 1176 
hos, 1176 
bosa, 1176 

iv, 118a 

leSer, 118a 

Old Slavonic Index. 


locian, 118a 
lus, 118a 

nest, 114a 
nefa, Ilia 
nift, 111a 

segel, 1186 
snegel, 1186 
sumor, sumer, 1186 
surma, 1186 

to, 1156 
tun, 1176 

vagen, 1186 
volcen, 1196 
vudu, 119a 


dark, 1176 

flat, note 87, p. Ill 
floor, 118a 

hat, 114 
herring, 114 
hog, 118a 

lame, 118a 

nut, 110a 

reader, 114 

smoke, 115 
strike, 115 
string, 114 

strive, 115 
swain, 115 

town, 1176 

witness, 119a 

Old Norse. 

ala, 1136 

baegjask, 117a 
baga, 117a 
bagi, 117a 
bagr, 117a 
bana, 1136 
bani, 113b 
batr, 117a 
belgr, 117a 
bogi, 117a 
bokki, 117a 
borS, 117a 

coss (koss), 1176 

dockr, 1176 
drott, pi. drottir, 1156 
drottin, 1156 
drottning, 1156 

egg, note 88, p. 112 
epli, 115a 

flatr, note 87, p. Ill 

glas, 1176 
griot, 1176 

hafr, 114a 

heill, note 84, p. 108 

iarl, 115 

iarn, v. isarn, 118a 

ior, 1106 

isarn, 118a 

kerra, 1176 
kreftr, 1176 

ledr, 118a 
lus, 118a 

naktr (nakinn), 116a 
nift, 114a 

segl, 1186 
snigil, 1186 
snara, 1186 
son (sonr), 121 
sumar, 1186 
sunna, 1186 
svefn, 1116 
svein, 115 

tivar, 1106 
tonn, 110a 
tun, 1176 

Jjak, 1126 
j?ior, 1096 

vagn, 1186 
vi£r, 118a 

yr, 118a 


Old Slavonic. 

ablani, abloni, v. jablani, 

agne, v. jagne, 1 116 
agnica, Ch. SI, 1116 
agnici, v. jagnici, 1116 
aice, v. jaice, 1126 
anigelu, Ch. SI, 114 
apostolu, Ch. SI, 114 
-ari, 124 
azu, 109a 
aza, 109a 

bada, 128 
brada, 996 
bratru, 1096 

A. Slavonic. 
bratu, 1096 

czrivi, 110a 
czrtiminu, 110a 
czruvi, 110a 

dehti, 115a 
deni, v. dini, 121 
desmti, 110a 
dmi, 110a, 121 
do-, 1156 
drevo, 110a 
drugti, 1156 
druva, 110a 
dvM, pi. 1106 

* geravjas, v. zeravll, 113a 

gnezdo, 114a 
golctbi, 1006 
govedo, 1096 
greba, 115a 
grobu, 115a 

ida, 128 

igo (jigo), Ilia 

ime, 109a 

jablani, 115a 
jabltlko, jabluka, 115a 
jada, 128 

jagne, v. agne, 1116 
jagnici", v. agnici, 1116 
jaice, v. aice, 1126 
jaza, 109a 


Indices Verborum to Position of the Celtic. 

jeza, 109a 
junu, Ilia 

kobyla, 110a 
kobylica, 110a 
koleno, 1196 
koni, 110a 
krivti, 1176 
krticzagu, 115a 
krticziminica, 115a 
krucziviuiku, 115a 

ligiikti, Ilia 
liza, Ilia 

moga, 1156, 123 
mati, Ilia 
meseci, 1116 
mesiti, 1116 
innogu, 116a 

nagu, 116a 
nebo, 119a 

ogni, 119 
onti, 127 
orllu, 1156 


angelas, 114 
anksztas, 109a 
ankszta, 109a 
ans, 127 
apasztalas, 114 
aszara, 110a 
aszva, 1106 
at-, 113a 
auksas, 996 

barzda, 996 
brolis, 1096 

da-, 1156 
dalis, 115a 
dalyti, 115a 
dantis, 110a 
deni, 121; 
dena, 110a 
derva, 110a 
deszine, 110a 
devas, 110a 

draugalas, 1156 
draugas, 1156 
durys, 1106 

pameti, Ilia 
plunti, Ilia 
prijateli, 116a 

sladukii, note 82, p. 97, 124 

sejati, 1166 

sesti, 112a 

sestra, 112a 

sladuku, note 82, p. 97, 124 

slunice. 112a 

snegu, 1166 

snocha v. snucha, 121 

snucba, 121 

sobaka, 110a 

soli, 1116 

srudice, 1096 

struja, 1116 

struga, v. struja, 1116 

sucbu, Ch. SL, 112a 

suka, 110a 

sunti, 1116 

svekru, 1116 

svekruvi, 1116 

svekry, v. svekru, 1116 

tima, 1126 
tiniku, 1126 
trtinu, 1156 
turti, 1096 

B. Lettic. 

erelis, 1156 
eris, 1156 
-esnis, 126 

gale ti, 119a 
galiu, 119a 
gandras, 1156 
gelbeti, 123 
gerve, 113a 
gimti, 1106 
grabas, 115a 
gyvas, 1096 

inkaras, 99a 

jaunas, Ilia 
-jaus, 126 
-jausei, 126 
-jausias, 126 
jungas, Ilia 

kampas, 110a 
karczama, 115a 
kelys, 1196 
kirmele, 110a 
kirminas, 110a 
kirmis, 110a, 119a 
koravoti, 119a 

vazvi, 109a 
veczeru., 1126 
ridova, 1126 
visi, 1126 
vlada, 1166 
vladiti, 1166 
vlasti, 1166 
vluku, 120 
vltina, 1126 
vranu, 119a 

zena, 1096 
zenti, 1106 
zeravli, 113a 
zima, 1106 
zivti, 1096 
znaja, 1106 


jaje, 1126 

wart, 1166 
wilk, 120 


junak, Ilia 

kreivas, 1176 
kuinas, 110a 
kulnis, 1196 
kumele, 110a 
kumelukas, 110a 
kumpas, 110a 

laizau, Ilia 
lengvas, Ilia 

maiszyti, 1116 
melynas, 1196 
menes, v. menu, 1116 
menu, 1116 
merga, 1196 
mergele, 1196 
moketi, 116a 
m&ku, 116a, 123 
mote, Ilia 

nugas, 116a 

obelis, 115a 
obulas, 115a 
-orius, 124 

pilnas, Ilia 
pirm, 126 
pirmas, 126 

Old Irish In dew. 


plat us, Ilia 
pretelius, 116a 

ratas, 1116 

sakau, 112a 
saldus, note 82, p. 97 
sapnas, 1116 
saule, 112a 
saiisas, 112a 
sekla, 1166 
seku, 112a 
semens, 1166 
Senas, 112a 
senis, 112a 
sesti, 112a 
sesii, 112a 
Scti, 1166 
snegas, 1166 
snocha, 121 
snttcha, 12 L 
sraume, 1116 
stogas, 1126 
szu, 110a 

szirdis, 1096 

tamsa, 1126 
tauta, 1126 

ugnis, 119 

vakaras, 1126 
valdau, 1166 
valdyti, 1166 
varna, 119a 
varnas, 119a 
venas, 1116 
vertas, 1166 
vese'ti, 1126 
vesz-pats, 1126 
vilkas, 120 
vilna, 1126 
vyras, 1126 

zema, 1106 
zinau, 1106 


abols, 115a 

dallit, 115a 

debbes, 119a 

dl-ws, 110a 

draudse (*draugia), 1156 

erglis, 1156 

gows, 1096 

krogs, 115a 

sapnis, 1116 

tauta, 1126 

waldit, 1166 

wens, 1116 

wirs, 1126 

Old Prussian. 

ains, 1116 
dellieis, 115a 
eranes, 109a 
* ganna, 1096 
tauta, 1126 
werts, 1166 
widdewu, 1126 

Old Celtic. 

ad-, 113a 

ande-, 99a, 123 (St. 734) 
Argento-ratum, 996 (St. 

Arduenna, 1076 
ate, note 103,^. 113 

bulga, 117a (St. 217) 

Camba, 110a 
Cainbodunum, 110a. (St. 

p. 150) 
carrus, 1176 

cataracton, note 85, p. 108 
* catarax, note 85, p. 108 
Caturiges, 1176 
Catu-slogi, ! 1176(St. 1003) 
Cebenna, 1076 
covinus, (Brit., Belg.) 

Crixus, 1136 

dan (root), 122 
drungus, 115a 
dula, 122 (see 7rf/x7T£- 
SovXa ; St. 765) 

Gaesati, 1176 (St., Gai- 

sati, 216) 
gaesum, 1176 


Kdovov ri]V aakTriyya, 

Kdpvv%, 114a 

XayKta, 1086 
Lutetia, Luteva, 1086 

fiapKav, ace, 118a 
Mopacdfifit), 110a 

ve/xrjTOVj ace, 121 (St. 

ov^eWov, ovZeWa (Brit.), 

118a (v. St. 13J 
Oppianicnos, 123 

TrefnrkdovXa, see dula, 122 
(St. 765) 

'Seyojuapog, 121 (St. 423, 

p. 156) 
Seno-magus, 112a 

tarvos, 1096 (St. p. 159) 
Toutissicnos, 123 

Uxellodununi, 118a (St. 

vertragus, 107, 113a (St. 

vidu, 119a (St. 46) 

Old Irish. 

ab, 99a (C. I to.) 

aball, 115a (St. 555) 

abstanit, 99a 

accidit, v. aiccidit, 99a 

accus, v. ocus, 109a 

acbtail, 99a (C l w.) 

actegim, 99a 

acuit, 99a 

accus, v. ocus, ocuis, 109a 

achtail, 99a (C. Liv.) 

actegim, 99a 

acuit, 99a 

acus, v, ocus, ocuis, 109« 

ad-, 113a, 120 

adaltras, 99a (St. 882) 

adgensa, 1106 

adgeuin, 1106 

adiect, adiecbt, 99a 

admuinur, Ilia 

adrad, 99a (C. I. w.) 

adras, 99a 

adrorsat, 99a 

adsaitis, 112a 

aeclis, dat. abh; v. gen., 

ecolso, ecilse ; 1016 
aer, v. aiar, 107a 
ag (root), 107a 
agathar, 116a 

198 Indices Verborum to Position of the Celtic. 

aiar, v. aer, 117a 

-aib, v. -ib, 127 

aibgiter, 99a (C. /. w.) 

aiccent, aiccend, dat. aic- 
ciund, 99a 

aiccidit, v. accidit, 99a 

aicher, 99a ( C. I. w.) 

aichthi, 116a 

aidrech, 116a 

aile, 109a, 113 (St. 158) 

ailigirn, 107a 

aine, 103a (C. I. iv.) 

aingel (angel), 99a, 111 
(C. l.w.) 

ainm, 107a, 109a (St. 

airdircc, v. erdirc, irdirrcc. 

-aire, v. -ire, 124 

airecal, 104a 

airget, v. argat, 99a 

airlech, 103a 

airriu, 127 

ais (ois), gen. aisa, aisso 
(oissa, oesa), 109a (St. 

aith (*ati-), 113a (St. 155) 

aithgne, 1106 

aithirge, v. ithirge, 116a 

al, (root) 113a 

almsan, ace. almsin, 1016 

alt, 107a 
altoir, 99a 
altram, 113a 
-am, 126 

amail, arnal, 109a 
amlabar, 1186, 124 (St. 

amprom, amprome, 105a 
ana, 127 
ancretem, 124 
ancretmech, 124 
angel (aingel), 99 a, 114 
anim, 107a, 109 
apstal, 99a, 114 (C. I. w.) 
apstallacht, 124 
ar (root), 1096 
arafulsam, 113a 
arathar, 99a 
arbae, v. orpe, 1166 (St. 

752, p. 163) 
ardd, 1076 
arenindarbe, 117a (St. 

argat, v. airget, 99a 
arm, arm a, v. dat. fsind- 

airm, 996 
117a, 130 

art, 996 (C. I. w.) 
articol, gen. sing., nom. 

pi., articuil, dat. ar- 

tucol, 99-5 
arva (arba), 107a 
as (a, es), 1076 
asan, 996 

asdul, 1026 (astol, C. I. w.) 
asil, 113a 
asmecnugur, 124 
asrobrad, 130 
atbela, 1136 
athir, 1095, 113 (St. 13, 

athusu, 126 
atom, 995 
atomaig, 107a 
-atu, v. -etu, 124 
augtortas, 995 (St. 1107) 
augaist, 995 (C. I. iv.) 

bachall, 996 ((7. I. w.) 

badud, 117a 

baga, 117a 

bagim, 117a 

bagul, 117a 

baislic, 100a (C. I w.) 

baitsim, ace. baithis, dat. 

batbius, 996 
bal (root), 1136 
balb, 996 (C. l.w.) 
ball, 1076 (St. 638) 
ban, v. ben (root), 113a 
ban (mulier), v. ben, 1096 

(St. 21) 
bandechuin (pi.), 1016 
banscala, 1186 
bar, v. ber (root), 1096 
bare, 996 (C. l.w.) 
bas, 1136 
bathach, 1136 (St. p. 

bauptaist, 996 (C. l.w.) 
beisti,, 100a ( C. I. w.) 
bemen (pi.), 1136 
ben, v. ban (root), 113a 
ben (mulier), v. ban, 1096 
bendacbae, 100a 
bendacht, 100a (C. I. w.) 
beo, v. biu, 1096 
beod, 1096 
beogidir, 1096 
beotliu, bethu, 1096 
ber, v. bar (root), 1096 
berach, 1076 
bethe, 108a 
bethu, beotbu, 1096 
biad, 1096 (St ; 477) 
biail, biail, buail, 117a 
birdae, 1076 

biu, v. beo, 1096 

biu, 127 

bochaill, 108a (St. 583) 

boide, v. buide, 117a 

bolg, bole, 117a (St. 217) 

boll, 100a (St. 159 ; C I. 

borg (borcc), 1136, 117a 
borggde, v. borg, 1136 
bou, 108a, 1096 
brace, 100a (C. I. w) 
braich, 100a (C. I. w.) 
braisech, 100a 
bran, 119a 
brath, 122 (St. 336) 
breth, 122 (St. 336) 
brathair, brathir, 1096 (St. 

breib, 1006 
brithemnacht, 124 (St. 

broen, 117a 
buaid, 117a 

buail, v. biail, biail, 117a 
buide, v. boide, adj. 108a, 

(St. 803), subst. 117a 
buidecb, 117a 
buidnib, 117a 

caech, 1136 

cacht, 100a (C. I. w.) 

cailech, 100a 

caille, 104a, 114b (C. I. 

caimse, 100a (eaimmse, C. 

/. iv.) 
caindloir, 100a (St. 44) 
caingel, 100a 
caiptel, 100a (C. I. iv.) 
caire, 119a 
cairigud, 119a 
caisc, 104a (C. I. w.) 
caise, 1006 (C. l.w.) 
caisel, 1006 ( C. I. w.) 
calann, 103a (CI w.) 
callaid, 100a (C. I. w.) 
camm, 1096 
cammaib, dat. pi., v. 

camm, 1096 
cammderc, 1096 
camthuisil, 1096 
can (root), 108a 
canoin (ace), 100a (C. I. 

car (root), 108a 
carachtar, 1006 
carcar, gen. pi. carcre, 

dat. carcair, 100a 
carim, cairim, 128 
carmocol, 100a 

Old Irish Index. 


carpat, 1006 (C. I. w.) 

cast, 1006 

castoit, 1006 (C.Z.w.) 

cath, 1176 

cathir, cathair, 108a (St. 

cathlac, 1006 (C. I. w.) 
cathrach, gen. v. cathir, 

note 85, p. 108 
caut, 100a 
cedir, 1006 
ceir, 1006 (C. I. to.) 
ceirbsire, 1006 
ceist, 105a (cest, C. I w.) 
eel, 108a 

cell, 1006 (eel, C. I. w.) 
cen, 131 

cenaelugud, 1106 
cenel, 1106 (St. 676) 
cenelach, 1106 
cenelae, 1106 
cercenn, 1006 (C. I. w.) 
cerchaill, 1006 (C. /. w.) 
cercol (ace), 1006 
cetlaid, 108a (St. 3) 
cilic, 1006 
cimbal, 1016 
cingcidis, 1046 
cingices, 105a (cingciges, 

C. I. w.) 
circumflex, 1006 
ciuil (gen.), 108a 
cis, 1006 (C. /. w.) 
* cladibas, 108a 
claideb, 108a 
clais, 1006 
clechir, 1006 
clechti, 1006 
clerech, 1006 ( C. I. w.) 
climata (pi), 1006 ( c - '• 

clocc, 1006 
cloebmuer, 1006 
cl&i, 108a 

clum, 1046 (C. I. w.) 
cnam, 108a (St. 269) 
cira, 110a 

coehull, 101a (C. I. w.) 
coibse, 101a ( C. I. w.) 
coic, 101a, 122 (St. 776; 

C. I. w.'j 
c&is (dat.), 1006 (St. 434, 

ace sing.) 
coisecrad, 101 a (St. 880) 
colcaid, 1016 (C. l.w.) 
coll, 101a 
colomna (nom. pi.), 1006 

(C /. w.) 
colum, 1006 
columnat, 1006 

com-, 126 
comacus, 109a 
comadas, 101a 
comadasogod, 101a 
comaicsiu, 109a 
comalnadar, Ilia 
comarbus, 1166 
comarpe, 1166 
commescatar, 1116 
companacht, 101a 
comparit, pi. -\t\,gen. -ite, 

conflechtaigthi, 101a 
congnam, 1106 
conoscaigesiu, 1116 
conrobam, 131 
conrochra, 131 
conrogbaid, 131 
couroscaigissiu, 1116 
conson, gen. consine, 101a 
corcur, 105a 
corgais, 105a 
coro-, corro-, conro-, 131 
corp, 108a (C 7. w.) 
cos, 108a 
cosmail, cosmuil, cosmil, 

* consamali, 109a 
credal, 101a 

crepscuil, 101a (C. I. w) 
cresen, 1006 (C. I. w.) 
cretem, 124 
cretes, crettes, creites, pi. 

cretite, v. cretim, 108a 
cretim, 108a 
criad, 101a 
criathar, 108a (St. 700; 

(C. I. w.) 
cricbaib, 1136 
cride, 1096 (St. 1102) 
crismal, 1006 
crocann, crocenn (leg. 

croccan), 115a (St. 56) 
croch, 101a 

cruim, 110a, 1176, 119a 
cruimther, 105a (C. I. w.) 
cu, 110a 

cuach, 1006 (C. I. w.) 
cuicenn, 101a (cucenn, C. 

I w.) 
cuigel, 101a 
cuilennbocc, 117a (St. 

cuimlengaithi, 113a (St 

gl. N? 45, p. 147) 
ciiisil, 101a (C. I. w.) 
cul, 108a (C.l. w.) 
cumacc, 109a, 116a, 123 
cumacht, 109a, 116a 
cumachtach, 116a, 124 
cumacht(a)e, 116a 

cumacbtagimm, cumacht- 

aigim, 124 
cumachtchu, comp. v. 

cumachtach, 116a 
cumaing, 116a 
cum an (v. ni cuman lim), 

Ilia (St. in; 
cumang, 109a, 116a, 123 
cummasc, gen. cummisc 

cumsciget, 1116 
cumuing. v. cumaing, 116a 
cupris, 1016 
cusecar, 101a 
cute, 105a 

dairde, daurde, 110a (St. 

dam (root), 110a 
damilsi, Ilia 
dark (root), 113a 
daur, 110a (St. 554) 
daurauch, 110a (St. 554) 
daurde, dairde, 110a (St. 

deccu, 127 
demne, gen. pi. v. demuin, 

demuin, v. demne, (gen.), 

1016 (C. I. w., deman) 
* denge, v. tenge, 123 
denim, 112 (St. 899) 
der, 110a 

derucc, 110a (St. 554) 
der wen, 110a 
descipul, 1016 (deiscipuil, 

n. pi., C. I. w.) 
dess, 110a (leg. des, St. 

de't, 110a 
di, 1086, 120 
dia (dies), 110a (St. p. 

163 ; dia, C. I. w.) 
dia (Deus), 110a (St. 81 
diabul, 1016 
diblide, 1016 
dictatoir, 1016 
digaim, 1016 
dil, diliu, dilem, 115a (St. 

diles, 115a 
dilui, 1016, 115a 
dinair, 1016 

dipreeoit, 1016 (C. I. w.) 
discreit, 1016 (C. I. w.) 
do, du, 1156, 131 (St. 570) 
do-, v. du-, 1086 (St. 85) 
doaibsem, 127 
doaithirge, v. taidirge, 




Indices Verborum to Position of the Celtic. 

cloaurchanaim, 108a (St., 

doaurchanim, 704, 837) 
dodalim, 115a 
doforsat, 112a 
dofuibnimm, 113a 
dogentar, 1106 
dogniu, 1106, 127 (St. 908) 
doib,doib, 127 
doiseich, 112a 
-doit, 124 

domnach, 1016 (C I io.) 
domenarsa, Ilia 
domoiniur, Ilia 
donminursa, Ilia 
do-ornalgg, v. omalg, Ilia 
dorche, /. pi, 1176 (St. 

dor&sat, 112a 
dor us, 1106 
dosaig, 112a 
doseich, 112a 
drac, 1016 (C. I w.) 
draigen, 1156 (St. 559) 
driss, 1156 (St. 587) 
dristenach, 1156 (St. 587) 
droch, drog, 1146 
drochgnim, 1146 
drochgnimu, ace. pi, 1146 
drog, drocb, 1146 
droUhean, 1156 
du, do, 1156, 120, 131 (St. 

du-, v. do-, 1086, 120, 131 

(St. 85) 
duib, 127 
duibsi, 127 
duine, 122 (St. 89) 
dun, 1176 (St. 674) 
dur, 1016 

6, v. si, ed 

ech, 1106, 122 (St. 17) 

ecolso, ecilse, yen., v. 

aeclis, 1016 
ed, v. e, si, 126 
-em, 126 

emnatar, 1026 (St. 1010) 
eo, 118a 
epil, 1136 

epistil, 102a (C. I. w.) 
epscop, 102a (St. 982 ; C. 

erdaircigidir, 113a 
erdarcai, pi. v. airdircc, 

erdirc, v. airdircc, irdircc, 

erriu, erru, 127 
escalchaill, 102a (St. 115) 
estar, 1106 

etar, etir, v. itir, 1086 
etardibe, 1136 
etargeiuin, 1106 
etargne, etarcne, 1106 
etarru, 127 
ethemlagas, 102a 
etirdibnet, 113a 
-etu, v. -atu, 124 

fagde, 102a 

fallen, 1066 

faith, 1096 (St. 2) 

fedb, 107a, 1126 

fellsube, 1016 

felsub, 1046 (St. p. 159) 

fernin, 102a (femen, C. l^ 

fen, 1186 
fer, 1126, 121 
fers, gen. fersa, ferso, 1066 
ferte, nom. pi., v. ferto, 

ferto, ferte, gen, 1076 (C. 

I. IO.) 

fescor, 1126 (St. 224 ; C. 

I. w.) 
fetarlaice, fetarlice, fetar- 

licce, 107a 
fi, 1086 

nadnisse, 1186 (St. 959) 
fial, 1066 

fiar, v. sethar, 112a 
fich, 1126 
ficuldae, 102a 
fid, 119a 
fidbocc, 117a 
figil, 107a (C. I. io.) 
fimf, 122 

fin, f mn, 107a (C. /. tv.) 
fine, 107a (C. I. iv.) 
fir. 1066 
* firas, 121 
firaib, forib, 127 
firtu, ace., v. ferto, 1076 
fis, 1076 (CI. w.) 
fiaith, gen. Hatha, fiatho, 

fia(i)themnacht, 1166 
fiaithenmas, 1166 
fiur, 118a 

focul, 1075 (C. I. to.) 
fodail, fodil, 115a 
fodaimimse, 110a 
fodali, 115a, 122 
fodlaidi, 115a 
fognam, 1105(St. 815) 
fogni, 1106 
foircthe, 108a 
folcaim, folcaimm, 1196 

(St. 1045) 

fondrodil, 115a 
foraib, foirib, forib, 127 
forcanim,forchanim, 108a 
forchun, 108a, 127 
forcital, forcetal, 108a (St. 

forcitl(a)id, forcetlaid, 

108a (St. 837) 
forlan, Ilia 
forlongis, 103a 
forodil, 122 
forru, 127 

fuirib, v. foraib, 127 
fulang, 113a 

gab (roof), 114a 

gabail, 123 

gabor,114a, 1196, 123 (St. 

gabimm, 1196 
gabul, 1176 (St. 135) 
gaide, 1176 (St. 216) 
gaimred, 1106 
ged, 1156 

gein, gen. geine, 1106 
geinddae, 1106 
geinti, v. genti, 1026 
geinti, pi, 1026 (geinte, 

C. I. io.) 
geintlecte, gen. Jem., v. 

gentlide, 1026 
geintlide, 1026 
gell, 1176 

gen (root Skr. jan), 1106 
gen (root Skr. jna), 1106 
gen, dat. giun, 1136 
genitiu, 1026 
genti, v. geinti, 1026 
gentlide, 1026 
gentar, genthir, 1106 
gerind, 1026 
giaU, 1176 (St. 216) 
glass, etfas, 1176 (St., note, 

p. 91) 
gluais, 1026 
glun, 1196 
gne, 1106 
gnethid, 1106 
gnfm, 1106 (St. 908) 
gniu, 128 
grad, gen. graid, 1026 (C. 

I. io.) 
graif, 1025 

grazacham, 1025 (C I. w.) 
guidimm, 122 (St. 870) 

heritic, pi, 1026 
hiairn, gen. v. iarn, 118a 
ho buidnib, 117a 
hodid, gen. v. uathid, 1186 

Old Irish Index. 


(h)omaldoit, v. mnaldoit, 

horpamin, (pi.) v. orpam, 

hothad,t\ (h) uathath, etc. 

huathad, v. (h)uathath, 

etc.), 1186 
huathath ace., v. (h)ua- 

thath, etc. 1186 
huathati, fern. ace. pi., v. 

uaithed, 1186 
(h) umaldoit, v. umaldoit, 


iach, 102a (St. 216) 

iarm-, 126 

iarn, v. gen. hiairn, 118a 
(St. 608, 812) 

-ib, v. -aib, 127 

ibim, 1086 

id, 1026 

idol, 1026 (C. l.w.) 

ifurnn, gen. ifirnn, 103a 
(iffearn, iffern, St. 519) 

il,». lia, lllo(il, St. 13) 

irab-, 99a 

imdibe, 1136 

imdibenar, 113a 

immefolngai, immefoln- 
gai, immolhgai, 113a 

immeforling, imforling, 

-imem, 126 

immeruidbed, 1136, 130 

iramumruidbed, 130 

immolhgai (see immefol- 
ngai, etc.), 113a 

ind-, 99a, 123 (St. 734) 

indatbendachub, 100a, v. 

ind-figor, 102a 

indib, 127 

indid, 127 

indlach, 113a 

indlung, 113a 

infinit, 103a 

ingor, 99a (St. 68) 

inis, gen. inse, 1086 

init, 103a 

innarbar, 117a 

innerese (ace), 1026 

innoc(h)t, 1116 

innurid, 125 

inobar, v. saibes, 104a 

inroleg, 103a 

insadaim, 112a 

insce, 112a 

inte, 127 

interiecht, 103a 

intesi, 127 

intiu, 127 

-ire, v. -aire, 121 

irgnae, 1106 

isiad-airmm, dat. v. arm, 

isind-ithlaind dat., v. land, 

itargninim, 1106 
ith, gen. etha, 1086, 123 

(St. 1037) 
ithim, 1086, 1106, 1166, 

123 (St. 40) 
ithirge, v. aitbirge, 116a 
itir, v. etir, etar, 1086 
iugsuide, 103a 

labar, 1186 
lacht, 1086 
laech, 103a (C. I. w.) 
lagait, Ilia 

laigiu, lugu, Ilia (St. 
f 923) 

laine v. lane, Ilia 
Ian, Ilia (St. 13) 
lane v. laine, Ilia 
lanad, 122 
land, dat. isind-ithlaind, 

118a (St. 132) 
lang (root), 113a 
lanmair, Ilia 
lar, 118a 
lebor, v. libur, 103a (libar, 

St. 371 ; C. I. w.) 
lechdach, 103a (St. 1071) 
lecht, 103a ( C. I. w.) 
led, leth, Ilia 
legend, 103a (St. 853) 
legtoir, 103a 
* leic, 1086 
leicci, 1086 
leim, 118a 

leth, led, Ilia (St. p. 156) 
lethan, Ilia (St. 13,925) 
lethscripul, 1056 
li, 114a 
lia, v. il, Ilia 
liac, 1086 (liacc, St. 133, 

573, p. 156) 
libur, v. lebor, 103a 
ligim, Ilia 

lim (ne cuman lim), 111a 
lin, 103a (St. 863; C. I. 

Hnad, Ilia, 122 
linmaire, Ilia 
liter, 103a (letir, C. I. w.) 
loathar, 118a 
lobur, 1086 
loc, 103a 

loing, 103a 

loingtech, 113a 

long, 103a (St. 574 ; C. 

I. IV.) 

loth, gen. loithe, 1086 
luacharnn, 103a 
lugimem, Ilia (St. 923) 
lugu (see laigiu), Ilia 
luirech, 103a (St. 154) 
lunair, 103a 

mace, v. mang (root), 1156 

mace (Alius), 1 1 56 

maer (v. mor-maer), 103a 

(C. I. w.) 
magistir,, magis- 

tru, ace. pi., 103a (St. 

mainn, 1036 
maldachae, 103a 
maldacht, 103a (St. 915) 
maledic, 103a 
malg (root), Ilia 
man, 1036 
man (root), Ilia 
manach, 1036 (C. I. w.) 
mang (root, v. mace), 

mar (root), Ilia 
mar, 1156 (mar, St. 663) 
marb, Ilia (St. p. 159) 
marc, 118a 
martre, fern. pi. martri, 

1036 (St. 738; C.l.w.) 
masc (root), v. misc, 1116 
mascul, 1036 
matal, 1036 
naathir, Ilia 
matin, 1086 (C. I. w.) 
me, me, 126 
mebuir, 1036 
medon, Ilia 
meince, 116a 
meirddrech, 1036 
melim, Ilia 
membur (pi), 1036 
mencain, 116a 
menicc, menic, 116a 
mertrech, 1036 (C. I. w.) 
messa, v. mi-, 118a (St. 

metair, metir (gen.), 1036 

(metuir, C. I. w.) 
mi, 1116 (St. 1117) 
mi-, 118a 

mias, 103b (C. I. iv.) 
midiur-sa, 1086 
mil, 1086 
mil, 1036 
mile (fem.\ 1036 
'15 B 


Indices Verborum to Position of the Celtic. 

mindchichthiu, 1036 
mindchigitir, 1036 
mindeclm, 1036 
mirt-chaill, 1036 (St. 115) 
misc (root), v. masc, 1116 
mistae, 1116 (St. 1051) 
mo-, mu-, 131 
mod, v. muid, mud, 1036 
molt, 1036 
mong, 118a 

monistre (gen. pi), 1036 
m&r v. maer, 103a (St. 

moralus (dat.), 1036 
mori, Ilia 
moru, 1196 (mora, St. 

mu-, mo-, 131 
mucc, 118a (St. 1029) 
mud, dat, v. mod, 1036 
muid, gen., v. mod, 1036 
muin-torc, 1066 (St. 744) 
muir, 1116 (St. 860) 
muirtchenn, 1036 (C. I. 

mulenn, 1036 (St. 701 ; 

muilenn, C. I. w.) 
niulu (ace. pi), 1036 (St. 

mur, 1036 (C. I w.) 
mtit, 1036 

nachiberpidsi, 1166 
nachimrindarpai-se, 117a 
nat, 104a 

nathir, 114a (St. 88) 
naue (gen.), noe, 1086 
nebthobe, v. nephthobe, 

1136 (St. 987) 
necht, 114a (St. 224) 
* nem, 119a (St. 812) 

nephthobe v. nebthdbe, 

nert, 1086 
nessa, nesam, 114a (St 

neutor, 104a (neutur, C. 

I w) 
neutralde, 104a 
ni-, Ilia, 130 
niae, 114a 
nicumanlim v. ni, cuman, 

lim, Ilia 
nicumscaichti, 1116 
nid, 114a 

nimb, 104a (C. I w.) 
niroimdibed, v. roimdibed, 

riiule v. niulu, 113 
niulu, dat. (in niulu) 1116 

noacuitigfide v . acuit, 99a 
nobbendachat, 100a 
nobirpaid, 1166 
nocht-chenn, 116a 
noct, 1116 
nogigned, 1106 
nometargnigedar, 1106 
nomerpimm, 1166 
nomisligur, 124 
not,nota, 104a 
notail, 113 a 
notaire, notire, 104a 
notlaic, 1036 (C. I w.) 
nu-, no-, 131 

nu (nua), nue, nuae, nuide, 
1116 (St. 21, 803) 

oa, 1186 (St. 758) 
obar, 104a 
oblann, 104a 
ocbt, 122 
ochte, octe, 109a 
oclachdi, 1106 (St. 758) 
ocmil, Ilia (St. 758) 
ocus v. accus, 109a 
ocus, ocuis (e^) v. acus, 

og, 1126 (St. 955) 
6in, &en, 1116 
oipred, 104a (St. 889) 
oipretho, gen., v. oipred, 

oir, gen., v. or, 996 
oirclecb, 104a 
ois, v. ais, 109a (St. 

oissa, oessa, gen., v. ais, 

ois, 109a 
oistreoir, 104a 
olachaill, 104 a 
olachrann, 104a 
(h)omaldoit, v. (h)umal- 

doit, 1026 
omalgg, (v. do omalgg), 

onoir, 1026 
ood, v. uad, 127 
optait, optit, 104a 
orait, 104a (oroit, C. I. 

ord (ordd, ort, urt), 104a 

(St. 943; C.l.w.) 
orpam, v. pi h&rpainin, 

orpe, v. arpae, 1166 
orthain, ace. sing., 104a 

(C. I w.) 
ort, v. ord, 104a 
6s, v. uas, ucb, 118a 
othatnat, 1186 

othud, dat., v. uathuth, 

pagan, 104a 

pain, 104a (C. I. w.) 

paiper, 104a 

pairche, 104a (C. I. w.) 

pairt, 104a (C. I. w.) 

pais, 1046 (C. I. w.) 

papa, 104a 

partus, 104a 

peccad, 1046 (C. I w.) 

pellec, 1046 

pen, v. pian, 1046 

peneult, 1046 

pennit, 1046 (pennait, C. 

1. w.) 
persan, 1046 (St. 87) 
pian, v. pen, 1046 
piss, 1046 (C. l.w.) 
plag, 1046 
popul, 105a 
port, 105a (St. 676, 725 ; 

C. /. w.) 
posit, 105a 
predacb, 105a 
predchim, 105a 
precept, 105a (C. I. w.) 
preceptoir, 105a 
predag, 105a 
prelait, 105a 
prim, 105a (C. I. w.) 
proind, 105a (C. I. w.) 
prolach, 105 a 
promfldir, 105 a 
pronomen, 105a 
propost, 105a (C. I. w.) 
pupall, 104a 
pur, 105a ( C. I. w.) 
purgatoir, 105a 

ra-, v. ru-, ro-, 130, 131 

rad, 1166 

ram, 1056 

ranglana, 130 

rastal, 105a 

rect, reebt, 1146 

reilic, 1056 (relic, C. I. 

remi-, 126 
ri, 114a 
riagul, riagol, 1056 (St. 

ribar, 101a (C. I. to.) 
rig, gen., v. ri, 114a (St. 

ro-, v. ra-, ru-, 130, 131 
roainmnichte, 130 
roairptha, pi. v. rcerbad, 


Old Irish Index. 


robeirnmis, 130 
robia, robbia, ropia, 130 
rochumscigther, 1116 
rocomalnither, 130 
roerbad, 1166 
rofetar 123 
rogad, 122 
rogen(a)ir, 1106 
roiccu, 127 
roimdibed, 130 
rolin, Ilia 
rolabrastar, 128 
rommunus, rommunus, 

rondpromson, 105a 
ronoibad, 130 
ropia, v. robia, 130 
rorelus, 130 
roschaill, 1056 
rostae, 1056 
rostan, 1056 
rotb, 1116 

ru-, v. ra-, ro-, 130, 131 
rucestaigser, 128 
run, 1186 

sabaltair, 1056 (C. I. w.) 
saboit, 1056 (C. /. w.) 
sacardd, 1056 (sacart, C. 

sacc, 1056 
sacorbaic, sacarbaic, 1056 

(C. /. w.) 
sad (root), 112a 
sai, 109a 
saibes (saibes inobar), v. 

inobar, 104a 
saichdetu, 112a 
saiged, 112a 
saiges, 112a 
saiget, 1086 (St. 214) 
saigid, saiged, 112a 
saigim, 112a 
saigul, 1056 (St. p. 146) 
saile, 1056 (St. 651; C. 

sailestar, 1056 (soilestar, 

C. I. w.) 
* saillim, 109a 
sailm, pi., v. salm, 105a 
sak (root — to say), 112a 
sak (root — to follow), 

salami, 1116 (St. 977) 
salm, 105a (C. I. w.) 
salmu, ace, v. salm, 105a 
salt, 1056 [C. I. w.) 
saltair, 105a 

saltir, dot., v. saltair, 105a 
salto, gen., v. salt, 1056 

saltrach, gen., v. saltair, 

sam, 1186 
samail, samal, * samali, 

* samali, 109a 
sancht, 1056 (St. p. 161 ; 

C. I w.) 
sapati, pi., v. saboit, 1056 
scath, 1136 
sciath, 109a 
scipar, 1046 
scol, 1056 (St. 338) 
scoloca, 1186 
scribend, 1056 (St. 853) 
serin, 1056 (C. 1. w.) 
scnle, gen., v. scol, 1056 
sech, 109a 
sechem, 112a 
sechimtid, 112a 
sechtmaine, 1056 
seib, 102a (C. I. w.; St. 

seinser, 1056 (seindser, C. 

I. IV.) 

seit, dat, v. s6t, 1186 
seitchi, dat., v. setche, 

sen, 106a (C. I. w.) 
sen, 112a (St. 735) 
senatoir, 1056 
sens, 1056 
seol, s&ol, 1186 
septien, 106a 
set, 1186 (St. 470, 1073) 
setche, 1186 
sethar, siur (siar, fiar), 

108a, 112a 
seuit, seuit, pi., v. set, 

si, v. e, ed, 126 
siansib, dat. pi. v. sens, 

siar, v. sethar, 112a 
sfd, 1136 
sil, 1166 
sillab, 106a 
siur, v. sethar, 112a (St. 

slechtaim, 102a 
slice, 1186 
slid, pi. 1186 
snathe, 1186 (St. 817) 
snathiu, dat., v. snathe, 

snechti, 1166 
soillse, 112a 
sool, v. seol, 1186 
sosad, sossad, 112a 
spiracul, 106a (C I. w.) 
spirut, 106a (C. /. w.) 

sponge, 106a ( C. I. w.) 

srathar, 106a (St. 262) 

srian, 1026 (St. 109, 1039) 

srogell, 102a 

sruth, 1116 (St. 999) 

stan, 106a 

su-, 1096, 120 

suan, 1116 

such, 1186 

suide, 112a (St. 812) 

suidiguth, suidigud, 112a 

suist, 1026 (sust, C. I. w.) 

superlait, superlit, pi. 

superlati, 106a 
surnn, 1026 

ta, 127 

tablaire, 106a 

taidirge, v. doaitbirge, 

taig, dat., v. teg, 112a 
taigae (idultaigae, gen., v. 

teg) 112a 
* taigi, v. teg, 112a 
tairm-, 126 
talland, 106a 
tarn, 106a (C. I. w.) 
tana, 1126 (St. 1017) 
tar, 1096, 120 
tarb, 1096 
tau, 127 
tech, v. teg, 112a (St. 

teg, v. tech, 112a 
teirt, 1066 (C. I. w.) 
teis, 1066 
tellrach, gen., v. telluir, 

telluir, 106a (tellur, C. 

I. w.) 
temel, 1126 
tempul, 1066 
tene, 119 

tengad, pi., v. tenge, 1146 
tenge, gen. sing., 1146, 123 
teoir, 1066 (St. 744; C. 

tercital, 108a 
tesc, 1016 * 
test, 1066 (teist, C. I. w.) 
testimin, 1066 (testimon, 

testimoin, C. I. w.) 
tiach, 1066 (St. 41, 371) 
tiagu, 127 

tigerne, dat. tigerni, 1126 
timpan, 1066 
tir, 1096 (St. 703) 
titlu, ace. pi., v. titul, 1066 
titul, titol, 1066 
tobe, 1136 


Indices Verborum to Position of the Celtic. 

togu, 127 

tort, 1066 (C. I. w.) 

tot-mael, 1066 

tracht, 1066 

trag (root), 113a (St. 74) 

traig, 113a (St. 71) 

tre, tri, 1186, 120 

trebun-suide, 1066 

tremi-, 126 

tri, v. tre, 1186 

trindoit, 1066 

til, tu, 126 

tuath, 1126 (St. 423) 

tucu, tuccu, 127 (St. p. 

tuib, gen., 1066 
tuinech, 1066 (C. I. to.) 
tuir, 1066 
tus-lestar, 1066 (St. 1131) 

uad, ood, 127 

uadi(/e?«.), 127 

uadib, uaidib, dat., v. uad, 

liaithed, 1186 
uan, 1096 
uar, 1026 
uas, v. 6s, 118a 
uathataib,, v. liaith- 

ed, 1186 
uathath, uathad, v. (h)ua- 

thath, 1186 
uathid (hodid), gen., v. 

(h)uathath, 1186 
uathuth, dat, v. 6tkud, 

uch, v. 6s, and uas, 118a 
uile, 1186 

uilt, ace, v. ult, 1066 
ult, 1066 
umal, 1026 
(h) umaldoit, v. hunialdoit 

ungae, unga, 1066 
urde, 1096 
urt, v. ord, 101a 

*vlati, 1166 

Middle Irish. 

aibherseoir, 99a (St. 517) 

banprioir, 105a (St. 23) 
biait, 100a 

cabellanacht, 100a (cabil- 

lanacht, St. 172) 
command, 101a 
coroin, 101a (St. 75) 

crisdal, 101 

fairche, 104a (C. Lie.) 
fersaid, 1066 (St. 568) 
firmaniint, 102a (St. 749) 

gredail, 1026 (St. 854) 

instrumint, 103a 

orc(?), 105a (C. I.w.) 

proiste, 100a (St. 852) 
pusta, 106a 

sdair, 1026 (St. 84) 
seuadh, 106a (St. 551 ; 

senod, C. I. w.) 
sinistir, 102a 
sitheal, 106a (St. 241 ; si- 

thil, C. I ic.) 
soifist, 106a (St. 842) 
soiler, 106a (St. 740) 
spin, 106a 
stanamhail, 106a (St, 610) 

taibkerne, 106a (St. 169) 
tital, 1066 (C. I. ic.) 

Modern Irish. 

astaig, 1126 
meilg, Ilia 
pit, 1046 


aball, 115a 

aballea, 115a 

abbadeu,/>/. 99a 

abl, 1026 

acen, 99a 

agos, 109a 

alarch, 107a 

all, v. allt, 99a 

alldut, v. pi. alltudion, 

allor, 99a 
allt, alt, all, 99a 
alltudion, pi., v. alldut, 

amherawdyr, 1026 
aniberodraeth, 1026 
amherodres, 1026 
ampriodaur, 116a 
angor, 99a 
aniueil, v. pi. anyueilyeit, 

aniueileit, anniuieleit, v. 

anyueilyeit, 99a 

anniueileit, v. aniueileit, 

anuab, 124 

anyueilyeit, pi. v. aniueil, 

aradr, 99a 

araut, 104a 

archescyb,/)/., 102a 

arfeu, v. arueu, 996 

ariant, v. aryant, 99a 

armel, 996 

arueu, v. arfeu, 996 

aryant, v. ariant, 996 

assen, 996 

aual, pi. aualeu, aueleu, 

auon. 1076 
auonyd, 1076 
-awd, -awt, 128 
awel, 1076 
awr, 1026 
awst, 995 
awyr. 107a 

bad,/)/, badeu, 117a 

bagl, 996 

bahell, v. buyall, 117a 

baraf, baryf, 996 

barg, 996 

bar-wn, 100a 

baryf, v. baraf, 996 

bathor, 100a 

bedeu, 108a 

bedyd, 996 

bendicetic, 100a 

bendith, 100a 

bereu, 1076 

blodeu, 117a 

bodin, pi. bodiniou, 117a 

bord, v. bwrd, 117a 

boutig, 108a 

brag, 100a 

braut, brawt, 1096, 116a 

breich, 100a, 108a 

brodyr, pi. v. braut, 1096 

buch, 108a 

bud, 117a 

budicaul, 117a 

budugawl, 117a 

buyall, v. bahall, 117a 

bwa, 11 7a 

bwl, 100a 

bwrd, v. bord, 117a 

bwystuil, 100a 

byd, 128 

bydaf, 112, 128 

bydin, 117a 

bydwn, 128 

byleynyeyt, pi. 101 a 

byrdeu, pi. v. bwrd, 117a 

byw, 1096 

bywyt, 1096 

Welsh Index. 


cadeir, 1005 
cadwyn, 1006 
caeth, 100a 
calamennou, 100a 
callaur, 100a 
cam, 110a 
cancher, 100a 
cann, 108a 
cannwyl, 100a 
car, v. earr, 1176 
carbwncl, 100a 
cardotta, 100a 
cared, 119a 
carr, v. car, 1176 
carrei, v. corruui, 101a 
cath, 1006 
cawg, 1006 
caws, 1006 
celeel, 1016 
cenitol, 1106 
cenitolaidou, 1106 
cepister, v. kebyster, 100a 
ceroenhou, 1006 
cestill, cestyll, 1006 
chwaer, chwioryd, pi., v. 

chwior, 112a 
chwior, 108a, 112a 
circhinn, 1006 
eledif, cledyf, 108a 
cloeu, pi, 108a 
coc, 101a 

coch, pi. cochyoD, 1006 
cogail, 101a 
coll, 101a 
colenn, 101a 
coronawc, 101a 
corruui, v. carrei, 101a 
craff, 1176 
craffu, 1176 
crauell, 115a 
creaticaul, 101 
crefft, 1176 
cret, 108a 
criched, 1136 
crochann, 115a 
cristawn, 1006 
cruitr, 108a 
crych, 1136 
cudyaw, 110a 
cultel, 1016 
cultir, 1016 
cussan, 1176 
cwlltor, 1016 
cwydaw, 108a 
cyllell, pi. cylleill, v. kyl 

leil, 1016 
cymhar, 101a 
cymsc, 1116 
cymun, 101a 
cyson, 101a 

dacrlon, 110a 

dagreu, pi. 110a 

danned, pi. 110a 

dar, v. pi. deri, 110a 

datkanu, 108a 

-daut, dawt, v. -taut, 

decum, degura, 1016 
dehou, deheu, 110a 
derwen, 110a 
desko, 1016 
di, 1086 
didaul, 115a 
diffenu, 1016 
diffrwyth, 1026 
dihu, v. diu (dies), 110a 
din, 1176 
disci, 1016 
disgymm, 1016 
dispeilaw, 106a 
dm (dies), v. dihu, dyw, 

dyd, 110a 
diu (Deus), v. dyhu, dyu, 

dyuu, duw, duhu, duo, 

doeth, 1016 
doethou, 1016 
doguomisuram, 1036 
dolur, 1016 
dor, pi. doreu, 1106 
doythion, doeth, 1016 
draen, 1156 
dreic,/>?. dreigeu, 1016 
drog, 115« 
drogn, 115a 
dros, v. trus, 1096 
drus, 1106 
dnvc, 1146 
drws, 1106 
drycket, 1146 
dryssien, 1156 
duhu, v. diu, 110a 
duo, v. diu, 1 10a 
dur, 1016 
duw, v. diu, 110a 
duyuaul, 110a 
dyd, v. diu (dies), 110a 
dyhu, v. diu (dies) 
dyrys, 1016 
dyscyl, disgyl, 1016 
dyu, dyuu, v. diu, 110a 
dyw, v. diu (Deus), 110a 

eccluis, 1016 

ed, v. yd, 121, 130, 131 

eglwys, 1016 

egr, 99a 

egwyddor, 99a 

ehawc, 102a 

elestr, 1056 

eliffeint, 102a 
elin, 1106, 113 
emendassant, 102a 
enw, 107a 

epscip, pi. v. escyb, 102a 
erekafael, 130 
erw, 107a • 

escolectaut, 105 
escoleycyon, v. pi. ysco- 

leigyon, 1056 
escyb, v. epscip, 102a 
eskemun, 102a 
eskenho, eskynho, 1056 
estauell, v. ystauell, 100a 
eur, 996 

ffa, 102a 
ff\iwyd, 102a 
ffenigl, 102a 
ffer, 1086 
ffiol, 1046 
main, 102a 
ffo, 1026 
ffol, 102a 
fforch, 1026 
fforest, 102a 
ffrowyll, 102a 
ffrwyn, 1026 
ffurf, 102a 
ffust, 1026 
ffustawd, 1026 
ffynnawn, 102a 
fin, 102a 
finnaun, 102a 
fos, foss, 1026 
fruinn, 1026 
frut, 1116 
fruyn, 1026 
frwt, v. frut, 1116 
fual, 102a 
funenneu,/?/., 1026 
funiou, pi., 1026 

gaem, 1106 

gafar, 114a 

gallaf, 119a 

ganet, 1106 

gauar, 114a 

gayaf, 1106 

gerthi, 1176 

glin, 1196 

gofyn, Ilia 

golchi, 1196 

gormes, pi. gormesseu, 

gormessoed, 118a 
gorsedua, 112a 
gratell, 101a 
grawys, 105a 
gref, 1026 
grefiat, 1026 


Indices Verborum to Position of the Celtic. 

griduan, 1026 
gryd, 1026 
grydiaw, 1026 
gulan, 1126 
guletic, 1166 
guneyr, 1106 
guodeimisauch, 110a 
gwain, 1066 
gweddw, 107a 
gwennwyn, 1066 
gwenwynic, 1066 
gwerth, 1166 
gwerthawr, 1166 
gwerthyd, 1066 
gwiberot, 1076 
gwir, 1066 
gwlat, pi. gwladoed, 

gwledyd, 1166 
gwledic, 1166 
gwledyd, pi, v. gwlat 

gwnaf, 1106 
gwr, 1126 
gwrach, 1086 
gwyllt, 1196 
gwyrd, 1096 
gwystyl, 1176 

haf, 1186 

haliw, 1056 

ham, 1186 

he-, 1096 

heb, 112a 

hedwch, 1136 

helic, 109a 

helym, helm, 115 

henoid, 1116 

hestaur, pi. hestoriou, 

hestawr, 106a 
hestoriou, v. hestaur, 106a 
heu, 1166 
heul, 112a 
hewyt, 1166 
heyrn, 118a 
hint, 1186 
histr, 104a 
hoedel, hoedyl, 1086 
ho(s)an, pi. hossaueu,1176 
hucc, 118a 
huil, 1186 
huu, 1116 
hwylbrenni, 115 a 
hwyr, 106a 

iarll, 115 

ieuhaf, 1106 

ieuanc, pi. ieueinc, 1106 
inis, v. ynys, 1086 
iot, 1106 

iou, Ilia 

istrat, v. strat, 106a 

kaerllion, 103a 
kagbellaur, kyghellaur, 

kalaned, pi., 1006 
kanu, 108a 
karchar, 100a 
karw, 108a 

* kassiau, v. keissaw, 1086 
kauacus, 109a 

kaus, 1006 

kebyster, v. eepister, pi. 

kebystreu, kebesteryeu, 

keffyl, 110a 
kegin, 101a 
keissaw, 1086 
keleuyn, 100a 
kenedel, kenedl, kenedyl, 

kerwyn, 1006 
keryd, 119a 

* kessiaw, v. keissaw, 

keuedac, 1106 
kiwtawt, 1()06 
kiwtawtwyr, 1006 
kolouen, 1006 
koveint, 101a 
krissant, 101a 
kuyr, kwyr, 1006 
kyfagos, 109 a 
kyffelyp, kyffelyb, 109a 
kyfoeth, kyuoeth, 116a 
kyghellaur, v. kaghellaur, 

kylleil, jdZ., v. cyllell, 1016 
kyngryfet, 1176 
kynnhaeaf, 1106 
kyrchu, 1006 
kyuoeth, v. kyfoeth, 116a 
kyuoethawc, 116a 

laethauc, 1086 
laiSver, 1086 
lammam, 118a 
lann, 118a 
laubael, 117a 
laur, v. llawr, 118a 
leeces, 103a 
lemenic, 118a 
litan, Ilia 
huou, ph, 103 a 
llamp, 1 1 8a 
llanw, Ilia 
llawr, v. laur, 118a 
llawn, Ilia 
lie, 103a 

llech, 1086 
llegest, 103a 
Uei, Ilia 

lleidr, lleidyr, 103a 
llemhidyd, 118a 
Hew, 103 a 
lleycyon, pi, 103a 
lliw, 114a 

llong, pi. llorigeu, 103a 
lludedic, 1086 
Uwfr, 1086 
llygat, 118a 
llyghes, llynghes, 103a 
llygorn, 103 a 
llythyren, 103a 
loggeu, \oggou,pl. v. llong, 

maer, 103a 
mair, 103 a 
mal, 109a 
manaches, 1036 
maru, Ilia 
medeginyaethu, 1036 
medhecynyaet, 1036 
medic, 1036 
medwl, medol, 1086 
medylyaw, 1086 
meitin (yr meitin), 1086 
meldicetic, 103a 
melen, 1196 
melin, melyn, 1196 
melineu, 1036 
melynyon, pi, v. melen, 

meneich, 1036 
menoent, Ilia 
merch, 1196 
mererit, 1036 
metrut, 1036 
meun, Ilia 
milinon, pi, v. melen, pi. 

melynyon, 1196 
milwr, 1036 
mis, 1116 
moch, 118a 
modreped, Ilia 
mogau, pi (leg. moggou), 

morthol, 1036 
morwyn, 1196 
mur, pi. muroed, 1036 
mut, 1036 
mwng, 118a 
mwys, 1036 
mynnir, Ilia 
mynych, mynnych. Ilia, 

mynyd, 1086 
myrd, 1036 

Welsh Index. 


myrthw, 1036 
niys, v. mis, 1116 
mywn, Ilia 

nadolic, v. nodolyc, 1036 
neges, 104a 
neithawr, 104a 
nifer, v. niuer, 104a 
nimer, 104a 
niuer, v. nifer, 104a 
nodolyc, v. nadolic, 1036 
nouel, 104a 

oedawc, v. oetawc, 1086 

oedwn, 128 

oen, 1096 

oes, 109a 

oet, 1086 

oetawc, v. oedawc, 1086 

ofrum, 104a 

ois, 109a 

oleu, 104a 

padell, v. patel, 1046 

palf, 104a, 113 

pall, 104a 

pap,/?Z. papeu, 104a 

parchell, 105a 

part, parth, pard, 104a 

pasc, 104a 

patel, 1046 

pau, 104a 

pawin, 1046 

pawl, v. polyon, 104a 

pebreid, v. pybreid, 1046 

pebyll, 104a 

pechaut, 1046 

pedestyr, 1046 

pelechi, 1046 

penydyaw, 1046 

pererin, 1016 

peri, 104a 

person, 1046, 116a 

peth, 1046 

petbedic, 1046 

piumauc, 1046 

plycca, 1046 

poguisma, 1046 

polyon,/*?., v. pawl, 104a 

pont, 105a 

popuryes,pophuryes, r 101a 

porffor, 105a 

porth, v. pyrth, 105a 

portbant, 105a 

porthes, 105a 

porthi, 105a 

portbmon, 105a 

post, 105a 

postoloin, 105 a 

poues, 1046 

priawt, 116a 

prif. 105a 

prud, 105a 

pryf, 110a 

pull, 104a 

punt, 105a 

pur, 105 a 

purdu, 105a 

purgoch, 105 a 

purwynn, 105a 

putein, 105a 

pwys, 1016, 106a 

pybreid, v. pebreid, 1046 

pyllawc, 104a 

pyrth, pi, v. porth, 105a 

pysg, 1046 

pysgadwr, 1046 

rascl, 105a 

re-, v. ry, 130 

rebriuasei, 130 

reith, reyth,reis, 1146 

rhwyd, 1056 

rogulipias, 130 

rud, 1116 

rygaffel, 130 

ryt yssu, 1106 

ry (ry echewit), v. re-, 130 

saeth, 1086 

sant, v. pi. seint, 1056 

sarff, 109a 

scribl, 1056 

seint, pi., v. sant, 1056 


stebill, pi., 106a 

strat, v. istrat, ystrat, 1 C6a 

strotur, 106a 

strouis, 106a 

suh, 1186 

sumpl, 106a 

swllt, 106a 

swyf, 1056 

sych, 1116 

taguel, v.taVel, 1146 

talu, 115 

tarater, taradyr, 1096 

taru, 1096 

-taut, v. -daut, 124 

taw, 1146 

tawel, v. taguel, 1146 

tei, pi, v. ty, 112a 

teml, 1066 

temperam, 1066 

teneu, 1126 

testion, 1066 

testu, 1066 

tewi, 1146 

tigern, 1126 

torch, 1066 

torth, 1066 

traeth, 1066 

tribedd, 1066 

trintaut, 1066 

trist, 1066 

tristau, 1066 

tristit, tristyt, tristwch 7 

truch, 1066 
trus, v. dros, 1096 
turwf, v. twryf, 1066 
twr, v. tyreu, 1066 
twrneimeint, 1066 
twryf, twrwf, 1076 
ty, v. pi, tei, 112a 
tynnu, 1066 
tyreu, tyroed,/*/., v. twr, 

tywyll, 1126 

ucher, 1126 
ufern, uffern, 103 a 
uyeu, pi, 1126 

* vira, 1126 

wyf, 128 

y, 1086 

yd, v. ed, 121, 130, 131 

ychen, pi , 118a 

ymun, 1026 

ynyd, 103a 

yny (yny priawt person), 

ynys, 1086 

yscawl, v. pi, ysgolyon, 

yscolheic, v. escoleycyon, 

yscoleigyon, pi, v. esco- 
leycyon, 1056 

yscynnu, 1056 

ysgolyon, pi, v. yscawl, 

ysgymunn, 102a 

yskumunetic, 102a 

ysl(e)ipanu, 115 

ysmwg, 115 

ysnoden, 1186 

ysp, 1026 

yspeilaw, 106a 

yspeit, 106a 

yspwys, 106a 

ystabyl, 106a 

ystauell, 106a 

ystondard, 106a 

ystrat, v. strat, 106a 

yswein, pi. ysweinieit, 


Indices Verborum to Position of the Celtic. 


ad-, v. at-, 113a, 120 

all, 109a 

at-, v. ad-, 113a 

bran, 119a 

caer, cair, 108a 

* cardaut, 100a 

cat, 1176 

cav (i.e. cabh), Ilia 

cavael, 123 

ci, 110 a 

corf, 108a 

corn, Ilia 

dant, 110a 

do, dy, di, y, 1156, 120 

dyn, den, 122 

ech, 122 
-es, 124 
esel, 113a 

*garan, 113a 
gavar, 123 
guid, 119a 
guin, 107a 
*guint, 1146 
gwnathoed, 128 
gwybydy, 128 

hep, heb, 109a 
ben, 112a 
hi, 126 

lavar, 1186 
Hn, 103a 

mam, Ilia 

march, 118a 

melin, 1036 

mil, 1036 

moilin, v. muilin, 1196 

mor, 1116 

muilin, v. moilin, 1196 

nef, 119a 
nerth, 108a 
nos, 1116 

oil, 1186 

pimp, 122 

rin, 1186 

tan, 119 
tat, 1096 

-taut, 100a 
tir, 1096 
troit, 113a 
trui, 1186 
tut, 1126 

uchell, 118a 
un, 1116 

y, v. do, 1156 
yd, 121, 131 


abat, 99a 

abestely,/*/. v., apostol,99a 

aflauar, v. mab, 118-6, 124 

ail, v. eyll, el, 99a 

als, 99a 

altor, 99a 

anauhel, 1076 

ancar (ancora), 99a 

ancar (anachoreta), 99a 

anow, v. (h)anow, 107a 

apostol, v. abestely, 99a, 

archescop, 102a 
argans, arghans, 996 
arvow, 996 

ascient, v. guan, 1056 
asen, 996 
auain, 1026 
auallen, 115a 
auhel, 1076 
auon, 1076 
avel, 109a 
a veil, 115a 

barf, baref, 996 

bat, 100a 

batales, 100a 

bathor, 100a 

bay, 100a 

bedeven, 108a 

benen, benyn, pi. beney- 

nas, 1096 
benegis, 100a 
benenrid, 1096 
benenuat, 1096 
ber, 1 076 
bethaff, 112 
beu, 1096 
bewe, 1096 

bewnas, bewnans, 1096 
biu, 1096 
blamye, 100a 
blodon, 117a 
boch, 117a 
bom, v. bum, 1136 
box, 100a 

brag, 100a 

braud, v. broder, 1096 
brech, 100a, 108a 
broche, 100a 
broder, v. braud, 1096 
buch, 108a 
bugel, 108a 
bum, v. bom, 1136 
burges, 1136 

caltor, 100a 

cam, 110a 

caaihinsic, 110a, 1186 

can, 1 08a 

cancher, 100a 

cane, 108a 

canores, 108a 

cantuil, 100a 

cantulbren, 100a 

carogos, 109a 

caruu, 108a 

caul, 1006 

caws, v. cos, 1006 

ceroin, 1006 

chefuidoc, 116a 

cheniat, 108a 

clauster, v. cloister, 1006 

clear, 1006 

clethe, 108a 

cloch, 1006 

cloireg, 1006 

cloister, v. clauster, 1006 

cober, 1016 

coir, 1006 

colom, 1006 

colter, 1016 

collel, 1016 

colviden, 101a, 119a 

comfortye, 101a 

commisc, 1116 

corden, v. kerdyn, 1006 

cos, v. caws, 1006 

cothe, 108a 

creador, 101a 

cref, v. cryff, 1176 

croadur, 101a 

crogen, 115a 

croider, 108a 

crois, crows, 101a 

cruitr, 108a 

cryff, v. cref, 1176 

cugol, 101a 

cuic, 1136 

curun, 101a 

cussin, 1176 

cusul, cussyl, cusyl, cusill, 

cuthe, 110a 

dagrow, 110a 

Cornish Index. 


dampnye, 1016 

damp n ys, 1016 

dans, 110a 

dar, v. deri, 110a 

darat, 1106 

deserya, 1016 

det, 110a 

deth, v. dyth, 110a 

dethiow,/^., v. deth, dyth, 

dew (dew sull), v. sull, 

dewolow, pi, v. dy wolow, 

deyow, 1016 
diagon, 1016 
diffenor, 1016 
diseebel, v. dyscyplys, 

dyscyblon, 1016 
discomfortys, 101a 
diskient, 1056 
diskynna, 1016 
disliu, 114a 
doyn, 112 

drain, pi drein, 1156 
dreyn, pi v. drain, 1156 
dris, drys, 1096 
drocger, 1146 
drocgeriit, 1146 
drochoberor, 104a, 1146 
drok, 1146 
dug, 110a 
dyghow, 110a 
dyscyblon, pi, v. dis- 

dyscyplys, pi, v. diseebel, 

dyskas, 1016 
dysky, 1016 
dyspyth, 1016 
dyth, v. deth, 110a 
&ywa\ow,pl, v. dewolow, 


eddrek, 116a 
edrege, 116a 
eglos, 1016 
ehoc, 102a 
el, v. ail, eyll, 99a 
elerhc, 107a 
elin, 1106 
emperiz,/m. 1026 
emperur, 1026 
encois, 103a 
enef, 107a 
er, 1156 
ereu, 107a 
ermit, 102a 
erv, 107a 
escop, 102a 

estren, 104a 
eunhinsic, 1186 
eur, v. owr, 996 
eyll, v. ail, el, 99a 

fadic, 1026 

falhun, 102a 

fall, 102a 

fallens, 102a 

fellet, v. guin-fellet, 102a 

fenester, 102a 

fenochel, 102a 

feth, v. fyth, 102a 

ficbren, 102a 

finweth, 102 a 

fiol, 1046 

firmament, 102a 

flair, 1026 

flam, 102a 

fo, 1026 

fodic, 102a 

fol, foil, 202a 

forn, 1026 

frot, 1116 

funten, v, fynten, 102a 

fur, 1026 

furf, 102a 

fyll, 102a 

fynten, fynteon, v. funten, 

fyth, v. feth, 102a 

galloff, 119a 

garthou, 1 176 

gauar, 114a 

geaweil, 102a 

genys, 1106 

ghel, 1176 

gluan, 1126 

golhy, 1196 

gorthye, 1166 

gotheff, 110a 

gothevell, 110a 

govynny, Ilia 

goyf, 1106 

goyn, 1066 

graf, v. gwraff, 1106 

gras, 1026 

grat, 1026 

grauior, 115a 

grevye, 1026 

grou, 1176 

guan (guan ascient), 1056 

guedeu, 107a, 1126 

guein, 1066 

guenoinreiat, 1065 

* guid, 1156 

guil,1066, 1186 

guill, 1196 

guilter, 107 

guinfellet, 102a 

guins, 1146 

guir, v. gwyr, 1066 

guirt, 1096 

guistel, 1176 

guit, 1156 

gulat, 1166 

gur, v. priot, 112-5, 116a 

gurah, 1086 

gurthuper, 1126 

gustle, 1176 

gwerthe, 1166 

gwesper, 1126 

gwlas, 1166 

gwraff, v. graf, 1106 

gwyls, 1196 

gwyns, 1146 

gwyr, v. guir, 1066 

haf, 1186 

haloin, halein, 1116 

hanaf, 1026 

(h) anow, v. anow, 107« 

heligen, 109a 

hering, 114 

heuul, v. houl, 112a 

hiuen, 118a 

hoch, 118a 

hoirn, 118a 

hos, 1176 

hot, 114 

houl, v. heuul, 112a 

huir, v. piur, 112a 

huis, 109a 

hun, 1116 

huuel, 1026 

huueldot, 1026 

hveger, 1116 

hvigeren, 1116 

ieu, Ilia 
incoislester, 103a 
intre, v. yntre, 1086 
iouenc, v. jouonc, 1106 
iskel, 103a 

jouonc, v. iouenc, 1106 

kalagueli, 100a 

kat, 1006 

keghin, 101a 

kelegel, 100a 

kemeskis, kemyskis, 1116 

kerdyn, pi. 1006 

kigel, K) la 

kinethel, 1106 

kog, 10 1 a 

kyniaf, 1106 

lad, 103a 


Indices Verborum to Position of the Celtic. 

lacier, ladar ; pi. ladron, 

laddron, 103 a 
lagat, 118a 
lagas, 118a 
lait, 1086 
lear, v. ler, 118a 
legest, 103a 
leic, 103a 
len, v. leun, Ilia 
ler, v. lear, 118a 
leu, 103a 
leun, v. len, Ilia 
lilie, 103a 
litheren, 103a 
liuer, v. luffrow, 103a 
liuor, 114a 
logel, 103a 
lor, 118a 
loven, 118a 

luffrow, pi., v. lieur, 103a 
lugarn, 103a 
lyw, 114a 

mab, v. aflauar, 1186 
mair, 103a 
maister, 103a 
malou, 103a 
manach, 1036 
manaes, 1036 
mans, 1036 
mantel, 1036 
marth, 1036 
medhec, 1036 
medhecnaid, 1036 
mel, 1086 
menistror, 1036 
menit, meneth, 1086 
menough, 116a 
meras, 1036 
mesclen, 114 
mester, v. maister, 103a 
metin, 1086 
milin, 1196 
minfel, 1036 
mis, Ilia 
modereb, Ilia 
mols, 1036 
moroin, 1196 
moyrbren, 1036 
muis, 1036 
niynny, Ilia 

nader, 114a 
neid, 114a 
nevor, 104a 
noi. 114a 
noit, 114a 
noyth, 116a 

ober, 104a 

oberor, 104a 
ogas, 109a 
oin, 1096, 1116 
oleu, 104a 
oleubren, 104a 
olipbans, 102a 
or, 1026 
ors, 1066 
owr, 996 
oys, 109a 

padelhoern, 1046 

palf, 104a, 1116 

parchemin, 1046 

parth, 104a 

paun, 1046 

peber, 101a 

penakyll, 1046 

perbren, 1046 

pesadow, pi., 1045 

pesy, 1046 

peynys, 1046 

pinbren, 1046 

pirgirin, 1046 

pise, 1046, 114a 

piscadur, 1046 

piur, v. huir, 112a 

plentye, 1046 

plufoc, 1046 

plui, 1046 

pluuen, 1046 

pobel, pobyll, v. popel, 

pol, 104a 
pons, 105a 
popei, 101a 
popel, v. pobel, pobyll, 

porch el, 105a 
porth, 105a 
porthas, 105a 
porthow, 105a 
poruit, 104a 
pow, 104a 
powesough, 1046 
poys, 1046 
praysys, 105a 
prefis, previs, 105a 
prif, 110a 
prins, pryns, 105a 
princis, pi. v. prins, 105a 
priot (gur priot), v. gur, 

profuit, 105a 
profusy, pi. u.profuit,'105a 
pronteryon, pi. v. pro- 

under, 105a 
prounder, 105a 

redior, 114 

rethyskas, 130 
rewerthys, 130 
rewresse, 130 
roche, 114 
rud, 1116 
ruid, 1056 
ruif, 1056 
rute, 1056 
ruy, 114a 

sach, 1056 

saw, 1056 

sawye, 1056 

scauel. 1056 

scod, 1136 

scol, 1056 

scolheic, 1056 

screfe, 1056 

scriuen, 1056 

scriuiniat, 1.056 

scriuit, 1056 

scudel, scudell, 1056 

seithum, 1056 

sened, 106a 

seth, 1086 

settyas, 112a 

setva, 112a 

sibuit, 1056 

sicer, 106a 

skentyll, v. skyntyll, 1056 

skientoc, 1056 

skyntyll, 1056 

snod, 1186 

snoden, 1186 

soler, 106a 

sols, 106a 

spirit, 106a 

steuel, 106a 

stol, 106a 

strail, 106a 

strailelester, 106a 

streil, 106a 

strek, 115 

strekis, 115 

strevye, 115 

strifor, 115 

streing, 114 

stryff, 115 

suif, 1056 

sull (dew sull), 1016 

sur, 1056 

talgel, 1006 
tarow, 1096 
taw, 1146 
tempel, 1066 
temptye, 1166 
tenewen, 1126 
tensons, 1066 
termyn, 1066 

Armoric Index. 


tevolgow, v. tivulgow, 

the, 1086 
ti, 112a 
tist, 1066 
tistuin, 10G6 
tivulgow, v. tevolgow, 

tonnel, 1066 
trait, 1 066 
treason, 1066 
trech, 1066 
trest, 1066 
tribet, 1066 
trud, 1066 
tur, 1066 
tustunny, 1066 

ugh, 118a 
uncorn, 1066 
untye, 1066 
uy, 1126 

vertu, 1076 
vuludoc, 1166 

y-, 131 

yd, 1086, 121 
yffarn, yfforn, 103a 
yntre, v. intre, 1086 
yonk, 1'06 
yorch, 1086 


ael, v. el, hel, 99a 
ampeig, 1026 
aneualet, 99a 
apostol, 99a, 114 
appetaff, 1046 
applicquet, v. em, 99a 
archescob, 102a 
argant, 996 
aour, 996 

auber, v. ober, 104a 
auel, 1076 
auiel, 102a 

badez, 996 
ben(n)iguet, 100a 
bennoez, 100a 
beo, i096 
beuaf, 1096 
bezaff, 112 
bizif, 112 
brech, 108a 
breuder, pi., 1096 
bud, 1 17a 
buez, buhez, 1096 

cador, 1006 
cam, 1 10a 
christen, 1006 
clezef, 108a 
cloarec, 1006 
coar, 1006 
coffes, 101a 
coma(e)zreset, 101a 
comancc, 101a 
compizrien, 101a 
concedis, 101a 
coulm, 1006 
craf, 1176 
cref, creff, 1176 
cridif, 108a 
criff, 1176 
croeadur, 101a 
croeer, crouer, 101a 
cruel, 101a 
cusul, 101a 
cuzet, 110a 

dafnet, daffnet, 1016 

daffny, 1016 

decedy, 1016 

derch, 1 1 3a 

deiz v. diziou, 110a 

despez, 1016 

desquebl, 1016 

-det, 124 

di, 1086 

diaoul, 1016 

difen, 1016 

diner, 1016 

dimalicc, v. diualicc, 103a 

disguiblion, 1016 

disquennet, 1016 

disquif, 1016 

di.^prisonet, 105a 

diualicc, v. dimalicc, 103a 

diuiner, !016 

diziou, /?/., v. deiz, 110a 

doan, 112 

doe, 110a 

doen, 112 

doetaf, 1016 

doetanc, 1016 

douet. 1016 

douetaf, 1016 

drein,jo/., 1156 

dreist, 1096 

ed, 121 

el, v. ael, hel, 99a 

elin, 1106 

em (em em appliquet) 99a 

emtennet, 1066 

enef, 107a 

enes, 1086 

enterraf, 103a 

entre, 1086 
esper, 106a 
estonafF, 102a 
euel, 109a 
euffrou, pi. v. oberou, 

auber, 104a 
ez, 131 

falc'houn, 102a 

fall, 102a 

feunteun, feunten, pi. 

feuntenyon, 102a 
fez, feiz, 102a 
finisaf, 102a 
finuez, 102a 
flerius, 1026 
foil, 102a 
forest, 102a 
fos, foss, 1026 
foultr, 1026 
fruez, 1026 
fur, 1026 

gallaf, 119a 

ganet, 1106 

glin, 1196 

glisi, 1176 

gloan, 1126 

goestlas, 1175 

gouaff, 1106 

gousper, 1126 

gouzaf, gouzafl , 110a 

graf, v. groaf, 1106 

grif, v. groaff, 1106 

groaff, v. graf, gruif , griff, 

gruif, 1106 
guelchi, 1196 
guent, 1146 
guir, 1066 

haff, 1186 

baiarn, v. hoiarn, 118a 

hanu, 107a 

hastomp, 1026 

hel, v. ael, el, 99a 

hent, 1186 

histren, 104a 

hoiarn, v. haiarn, 118a 

bun, 1116 

iff am, 103a 
iusticc, 103a 

kaoter, 100a 
kemmeski, 1116 

lagat, 118a 
lann, 118a 


Indices Verborum to Position of the Celtic. 

lealtet, 103a 
lech, 103a 
leiffriou, 103a 
len, leun, Ilia 

mane, 1036 

manen, 1036 

martir, 1036 

maru, Ilia 

maruaill, 1036 

menaf, mennaf, v. minif, 

merch, 1196 
mesfectouryen, 1036 
mester, 103a 
rnillic, 103a 
minif, v. menaf, Ilia 
ministren, 1036 
mir, 1036 
miret, 1036 
miro, 1036 
musur, 1036 

nation, 104a 
nezaff, 1086 
nifer, v. niuer, 104a 
niuer, v. nifer, 104a 
noaz, 116a 
noter, 104a 

obediant, 104a 
ober, v. auber, 104a 
oberau, pi. v. ober, 104a 
oferen, pi. offer ennou, 

oben, 118a 
ordren, 104a 

paradis, paradoes, 104a 

parfetaff, 1046 

parz, v. perz, 104a 

pechet, pi. pecbedou, 1046 

peden, pi. pedennou, 1046 

penedour, 1046 

peoryen, pi., 1046 

perz, v. parz, 104a 

peuch, 1046 

pidif, pidiff, 1046 

plen, 1046 

pligadur, 1046 

ploeys, 1046 

ploi, v. plue, plueu, 1046 

plue, plueu, v. ploi, 1046 

poan, pi. poanyou, 1046 

pomell. 105a 

porz, 105a 

porzit, 105a 

pou, 104a 

poues, 1046 

prelat, 105a 

preservo, 105a 

prestis, 105a 

prezec, 105a 

priet, 116a 

proffe, prouffe, 105a 

psaulter, 105 a 

quercbat, querchit, 1006 

ra-, 130 

recommant, 1056 
reiz, v. rez, 1146 
renaff, 1056 
rento, 1056 
reol, 1056 

rez, v. reiz, 1146 
roed, 1056 
roen, 1056 

sacrileig, 1056 
saludomp, 1056 
sant, 1056 
sarmoner, 106a 
sceurt, 106a 
scler, 1006 
scoet-, scoit-, 109a 
sebeliaf, 1056 
sent, pi. v. sant, 1056 
sin, 106a 
soav, 1056 
soingaf, 106a 
soliter, 106a 
soutenet, 106a 
speret, 106a 
squient, 1056 
stat, 106a 

tardomp, 106a 
tempel, 1066 
teniff, 1066 
test, 1066 
ti, v. ty, 112a 
trindet, 1066 
tron, 1066 
ty, v. ti, 112a 

uasal, gen. v. us, 118a 
urz, 104a 
us, 118a 

vice, 1076 

ylis, 1016 


Gaulish and old Celtic. 

ande-, 143 
are-, 139, 164 
aremoricos, 165 
ate-, 139 
dunmo-, 139 
epo-, 139,161 
Esu-nertus, 139 

VEfXTJ-OV, 139 

Nerto-marus, 139 
Octodurus, 139 
Orgetorix, 139 
ver-, 162, 163 
vergo-bretus, 139 
Vernemetis, 139 


[The Modern Irish -words 
are printed in Italics, the Old 
and Middle Irish in Koman 

a, a, 159 

a, 160 

abair, 137 

abas, 165 

aca, 141, 155 

acaldam, accaldam, acal- 

tam, 143, 144, 145 
acarthar, 147, 160 
accomallte, acomaltae, 147 
accursagad, 159 
achesta, 159 
ad-, 137 
adarc, 143 
adbeir, 137 
adcomaltar, 147 
adgladathar, note, 144 
adgladur, note, 144 
adharc, 143 
adhradh, gen. adhartha, 

143, 149 
admuim, 145 
adrad, 143 
aecaillse, 139 
gecolsa, 139 
aedparthi, 135 
aes, 159 

agallam, 143, 145 
aggnim, 159 
aice, 141 
aichti, 147, 148 
aicnete, 149 

aid-, 139 

aidchuimthe, 147 

aige, 141 

aile, 140, 157 

ailiu, dat. masc, 157 

avnsear, 140 

aimser, 140 

ainmm, 140, 142, note, 

142, 158 
ainm, 158 
ainm, 140, 142 
air, 139, 165, 166 
air, 164 
air t note, 137 
airchinn, 165 
Airdeasbog, 159 
aire, 165 
airech, 165 
airi, note 138, 165 
airib, note 138 
airillti, 147 
airhim, note 138 
airiuibsi, note 138 
airiumm, 165 
airriu, note 138, 141 
airther, 165 
aiste, 141 
aith-, 137, 139 
aithdheanam, 150 
aithne, pi. aitheanta, 148 
aithte'idhte, 150 
alaile, 156 

alaili, gen. masc, 157 
alailiu, dat. 157 
Alba, 145 
Alban, gen. M.I. ; dat. 

Albain, 145 
Alpa, gen. Alpan, ace. Al- 

pai-ii, M. I., 145 
ambes, 159 
ambrotte, 149 
amires, 164 
amiressach, 164 
amprom, 162 
ammi, note 142 
an, 162 
an, 140 
a (n), 140 
an (n), 140 
an, 159 

anair, 140, 165 
anam, 142 
ananman, 159 
anasbertbar, 147 
anasbiursa, 160 

andeaghaidh, 143 
an dorogbid, 137 
angutas, 149 
aniendaB, 149 
animm, 142, note 142 
anfilaimm note 157 
aniiiambi, note 157 
anmammi, note 157 
anmanbi, 157 note 
anmande, 149 
anoir, 140 
ant, 140, 144 
aon, 140, 157, 158 
apair, M. I., 137 
apectba, 159 
Apilogdo, gen., note 139 
apir, 137 

apredchimnie, 160 
ar, 137, note 137, 166 
ar, 164, 165, 166 
ar', 165 
araii, 165 
arb (root), 139 
archenn, 165 
architinn, 165 
drd, 143, 146 
ardd, 143, 146 
argur, 165 
arinbretbre, 159 
arloure-ni, 159 
arri, 159 
arna, 147 
arnaib, 165 
arnet, 159 
arndiis, 159 
arnoib, 156 
arnoibbriathraib, 156 
aromfoimfea, 159 
arosailcther, 147 
arrocar, 160 
arsate, 149 
arse, 165 
arsid, gen. 149 
arsodain, 160 
ars&ire-ni, 159 
art, v. ardd, 146 
as, as, 165 
asagnintar, 147 
ascnam, 159 
asind-, 165 
asrirtber, 147 
asrulenta, 147 
ass, 165 
asta, 141 
ataimgt, 145 


Celtic Index to Phonology in Irish 

atbail, 144, 151 
atbeir, 137 
atdubelliub, 160 
athddidhte, 150 
athir, 138, 1(53 
athscribend, 151 
aththaoiseach, 150 
Atilogdo, gen, note, 139 
atobci, 160 
atobsegatsi, 160 
atomaig, 144 
aud-, 141 
aue, 141 

augtortas, 141, 145 
aiir-, 141 
aiirgabtha, 147 

ba, 161 

bade, pi. bailte, 143 

baill, 139 

baindea, M. I., 151 

baintigerna, M. I., 151 

baistim, 138 

baitsimm, 138 

banda, 149 

bandacblach, 151 

bandalem, 151 

bandea, 151, 159 

bandechum, 151 

banscala, 151 

bant&rismid, 151 

bard, 154 

barri, v. farii, 159 

baull, 139 

batillu, 139 

beag, 145 

be'arla, v. beurla, 138 

bee, becc, 145 

beirim, 137 

beisti, 135 

belre, 138 

bendachad, 143 

bendacht, 143 

benachadh, M. I., 143 

bennacbt, M. I., 143 

beos, 142 

ber {root) = Skr. bru, 137 

berli, 138 

berrthaid, 150 

berthir, 147 

bessti, bessti, v. beisti, 135 

betha, 139 

betho, gen. sing., 139 

beurla, 138 

bheirhn, 137 

bheos, 142 

biad, gen. biith, biid, 139 

biasta, M. I., 135 

biast, M. I., 135 

biddixnugud, 151 

bidh, 139 

bindd, bind, 143, 146 

binn, 143, 146 

bithgairddi, 149 

boill, 139 

bolad, 149 

bolgg, bole, 143, 146 

bolg, 143, 146 

boltigetar, 149 

borb, 143, 146 

borp, 143 

brage, 143 

bragha, 143 

breac, gen. brie, 145 

*brecc, 145 

breth, 139 

bnathraib,t\ arn&ibbriath- 

raib, 156 
brithem. 142 
brd. pi. brointe, 148 
buidh, 161 
brtide, 143 
buidhe, 143 
bullu, 139 
burbe, 146 
burpe, 146 

each, 137, 156, 157, 161 

caich, gen. v. each, 157,159 
caill, 140 

cailleach, 140, 143 
caillech, gen. caiilighe, 
^ 140, 143 
cain, 140 
calndloir, 143 
cainduthracht, 151 
caingnim, 157 
cainscel, 151 
caintaidlech, 151 
cainteist, 151 
caintoimtiu, 151 
caintol, 151 
caisc, 161 
caira, 140 
cairigthir, 147 
caoga 140 
caoin, 140 
caol, 140 
caora, 140 
carbad, 135, 145 
carbat, v. carput, 135 
carpat, M. I., 135, 145 
capuil, 154 
carput, 135 
cathrur, 157 
cead, 145, 158 
ce'adna, 143, 145 
ceann, 159 
cech, 156, 157, 159 
ceithre, 158 

ceo, 137 

cenalpande, 145, 149 
cene, 142 
ceneuil, gen., 157 
cenelu, ceneoll, ceneolu, 

ceniul, ceneoll, dat., 

cenn, cenn, 139, 150, 161 
cenodfil, 159 
centat, 150 
cedl, pi. ceolta, 148 
cesad, cesath, gen. cesta, 

cesto, 148, 159 
cessair, dot., 159 
cet, 145 

cethargarait, 149 
cethir, 161 
cethircliet, 158 
cetne, 145, 156, 162 
cetni, 157 
Chaisil, 159 
cheana, 142 
chenelaio', 157 
chesta, chesto, 158 
chetbutho, dat., 159 
chetni, dat. neut., 157 
cbetnai, gen., 157 
chetnidiu, dat., 157 
clilann, decl., 155 
clibi, dat., 157 
chluas, dat. chluais, 158 
choline-, 158 
cholnide, 157 
chos, 159 
chrann, 157 
chrann, 158 

chuca, ace. pi. 139, 141 
chugam, 139, 142 
chuice, 141 
chuige, 141 
chumachtig, 160 
cia, 137 
ciall, 158, 159 
cib e, cip e, 137 
cinmnfll, 160 
cinniud, 149 
claar, dat. 159 
claideb, 142 
claidhem, 142 
clainne, 140 
eland, 143, 161 
clann, 143 
clocc, 145 
clog, 145 
cloinne, 140 
cluain, p\. cluainte, 148 
cnoc, 145 
cnocc, 145 

co (ad, cum) 137, 142 
co (donee, ut) 137 

Celtic Index to Phonology in Irish. 


coaga, 140 
coatomsnassar, 159 
cocarti, 147 
codhladh, codladh, 143, 

cogadh, pi. cogtha, 149 
cofordumthesid-se, 159 
coic, 140, 161, note 161 
c6ica, 140 
coil, 140 

coill, pi. coillte, 140, 148 
coin-, 150 
coinneal, 143 
colann, 140 
colinn, 140 
colna, 143 
com', 165 

comalnad, note 144, 163 
comalnadar, 163 
comalnamar, 163 
comallnithe, 163 
comcbesad, 165 
comchlante, 147 
con, 160, 165 
con-, coin-, 150 
conde, 149 
condibfeil, 160 
condigcnte, 147 
condiuiti, 165 
condumfel, 159 
confesta, 147 
conintorgaitar, 147, 160 
connarcas, 143 
connaruchretesi, 147 
conrochretesi, 147 
consam, ace, 157 
conucbad, 144 
conulintae, 147 
coosnada, 165 
corcur, 161 
Cormac, note, 142 
coro, 137 
corp, 145, 157 
Corpimaqvas, note, 142 
cos, 165 
cosa, pi. 159 
coseitchi, 165 
cosmil, 165 
cosnadh, gen. cosanta, 148, 

cot-, 144 

cotaocbat, cotaucbat, 136 
cotlad, dat. cotlud, 145 
cotobsechf ider, 160 
cotondelcfam, 144 
crann, 161 
creidim, 145 
creitfess, ace. 157 
cren, 161 
cretim, 145 

creitme, 159 

cride, 143 

crldi, dat., 159 

croch, 139 

crocbad, gen. cr6chtho,139 

crocbtha, note 148 
crocbtbe, 148 
croidhe, 143 
crot, 145 
cruit, 145 
cruithnechta, 159 
cu, pi. cointe, 148, 149 
cue', 165 
cucci, 165 
cuccu, ace. pi, 139 
cuccunim, 139 
cuig, 140, 158 
cuing eis, 161 
cumactib, 165 
cundrad, 143 
cundradb, M. I., 143 
cunradh, gen. cunnartha, 

143, 149 
cursagad, gen. cursagtha, 

cutseltcbi, 159 

dd, 138, 158 

dagimrat, note 148 

dag- imrata, drog-imrato, 

dall-ciach, 150 
dam, 142, note 143 
damnae, note 136 
dan, 158 
danigud, gen. danigtbea, 

daonna, 149 

daoradh, gen. dadrtha, 149 
dare, (root) (— Gr. dspiuo), 

de., 153, 155, 156 
dealbh, 143 
deanadh, gen. deanta, 148, 

dearbh, note 143, note 145 
debtbach, 150 
debuitb, debthib, 

decbrugud, dat., 159 
deich, 158 
deirim, 137 
denti, 147 
dephtbigim, 150 
derbb, note 143 
derbtbair, 147 
desiu, 160 

dfaglibdil, d'fdgbhdil, 136 
d'fearthain, 150 

di, 141, 158, 159 
dia, gen. dei, de, 139, 152 
dian, 138 
dianaiper, 137 
dianeprem, 137 
dicheannaim, note 150 
diltuth, 150 
diguttai, dat., 157 
dirigutai, dat,, 157 
dintecnatatu, dat., 153 
dintrediu, dat., 153 
diobh, v. ddibh, 141, 155 
diombuan, note, 150 
diombuidheack, note, 1 50 
diomolaim, 150 
diofndr, note, 150 
diothoghluidhe, note, 150 
dlutai, 147 
do, 157, 158, 159 
do-, 137, 159 
do, HI, 153, 155,156, 159 
doair (d'air), 144 
doairci, 144 
doaith, 135, 144 
doaitbmne, note, 135 
doan, 138 
doaurchanim, 165 
doberrthe, 147 
dobimebomartt, 147 
dobinr, 137 
dobtromma, 160 
do-chantain, 150 

dofius, v. du-fius, 160 
dofoirde, 143 
dogentar, 147 
ddibh, v. diobh, 141, 155 
doilbtbid, 150 
doimin, 142 

doimmfolung, 158 
doinde, 149 

doinscannsom, 146, 151 
dom-, 142, note, 143 
* domain, 142 
dombersom, 137 
domthoscbid, 159 
dbmun, 139 
don, 157 
dond-, 158 
donebltar, 147 
dontorud, dat., 153 
doopir, 144 
doronta, 147 
dorurgabtha, 147 
dosceulaim, 151 
dosenmatbar, 159 
dosmbera, 160 
dothabairt, 159 
dothdgbhdil, 136 
drogimrato, 148 



Celtic Index to Phonology in Irish. 

drogscela, 151 
du-, 151 
dubhart, 137 

dubhras, 137 

dufius, v. do-fms, 160 
* dunad, 149 
dunattae, 149 
Dimpeleder, 151 
dus, duus, 160 

e-, 146 

€a, ei-, 150 

eadoimin, 146 

e'adtrom, 146 

eagcdir, 146 

eagna, 145 

e'an, decl. 155 

eardhairc, v. urdhairc, 

note 150 
eas-, 150 
easbha, gen. pi. easbhadh, 

note 143, note 145 
easbog, 138 
easpog, M.I., 138 
eatorra, 141 
ech, 139, 161 
ecintech, 149 
eclustai, 147 
ecne, 144, 145 
eceir, 146 
ed- 139 

edpart, v. idpart, 144 

eiblim, 145 
eidirj 141, 145 
eile, 140 
eiscsende, 149' 
eisserrgi, gen. 159 
eistim, 138 
eitsimm, 138 
en, 164 
epert, 148 
eperthe, 147 
epertbi, 147 
epertith, 149 
epll, 144 

epir, epiur, epur, 137 
*eplimm, 145 
epscop, 138 
er-, 139, 164 
Srbaid, 139 
erbid, 139 
erclioiliud, 148 
erend (tir-n-erend), 159 
. erriu, erru, note, 138 
erunn, note 138 
erutsu, note 138 
es-, 146 

esartaid, 149 

esib, 165 

etal, 139 

etar, 145 

etarru, 141 

etarscarad, gen. etarscar- 

tha, 148 
etha L 139 
etrum, 146 

facab, 136 
fad, 145 
fag, 136, 137 
fagbas, fagbus, 136 
fagebtis, 136 
fagk, 136, 137 
faghaim, 136, 137 
fagbbait, faghbat, M. I., 

fab", note, 137 
fait, 145 
farcluu, 159 
fardiull, dot. 157 
farn, v. barn, 159 
farnintbucbt, 159 
fear, 140 

fear, 140, decl. 155 
feara, voc. pi. y.fear, 157 
fearaibk, dat. y.fe'ar, 154 
feardha, 149 
fearg, 143, 146 
fedrr, 140 
Jealsam, 142 
felsnb, 139, 142 
fer, 140 
fer, 140 
fere, 143, 146 
ferce, dat., 159 
ferr, 140 
fescor, fescar, 161 

fointreb, 151 
foir, note 137, 163 
foirbtbe, 163 
foircheann, 163 
foirctbe, 147, note 147 
foirib, note 137 
foislte, v. bifoisite, dat. 

folnibtbe, 147 
fomam, dat. 159 
fomebridiebfider-sa, 159 
fomfirfidersa, 159 
fonsegar, 160 
for, 137, note 137, 157, 160, 

162, 163, 166 
foracab, 136 
foraib, forib, note 137 
forbanda, 147 
forcanim, 163 
forcanti, 147 
forcenn, 163 
forcetal, 163 
forcbain, 163 
forcbanim, 163 
forcbongair, forcongair, 

forcbongrim, forcongrini, 

forcbun, 163 
forcongur, 163 
fordomcbomaitber, 159 

fordubceebna, 160 

form, formm, note 137 

forndobcanar, 160 

forngarti, 147, 148 

forraind, note 138 

form, fornn, note 137 

form, note 137, 141 

forserce, 159 

fort, note 137 

fortbeit, forteit, 163, 165 

fds, 142 

fosmachtu, 146 

ficbe, gen. ficbet, 158, note fosodm, 160 

flche, gen. fichead, nom. 

fichid, 158 
fintan, 151 

fir, nom. pi. y.jear, 157 
fireanta, 149 
fo-, 137, 162 
foacbat, 136 
focbun, dat. masc. 157 
fodlaidi, dat., 157 
fogbaidefru, 136 
foghe'bha, 136 
foghlaim, 143 
foglaim, 143 
fogon, dat. 157 
foilsigud, 148 

fbt, 145 
fotracbussa, 136 
frecdairc, 143 
freendaire, dat., 157 
frecre, 144 
fri, 137, 160, 162 
fricacb, 165 
fris, 165 
fris, 165 
frisbiur, 165 
frisduntar, 147 
frislond, ace., 168 
friss, frissin-, 165 
fristacuirtber, 147 
frit-, fritt-, 151 
fritammiuratj 144 

Celtic Index to Phonology in Irish. 


frith', 162, 165 
frithaidechtge, 151 
frithcheist, 165 
frithtasgat, 151 
frituinthiagar, 159 
fudomain, fudumam, 142 
full, dat., 158 
/nil, 159 
fuiri, note 137 
fuirib, note 137 
ft&the, 141 
furastar, 147 
furib, note 137 
furnn, note 137 
futha, 141 

ga, 137 

gab, (root " capere"), 136 

gabail, 140 

gabar, 162 

gabal, 162 

gabhdil, 140 

gabhaim, 137 

gabimm-se, 137 

gach, 137 

gaith, gaid, 140 

^atafa, 149 

gan, 137, 154, 155 

gaoth, 140 

gentar, 147 

genthir, 147, 148 

gkeibhim, 137 

<7*"6e, 137 

go (to, with), 137 

</o (that), 137 

go-n-aibraim, 137 

grianda, 149 

grientairissem, 151 

guidimse, 152 

guv, 137 

gutte, gutae, 149 

hicc, 158 

hiclaar, v. claar, 159 
hifoMte, v. foisite, 159 
hipersonaib, v. persin, 156 
hires, 159 

hisintorunt, dat., 153 
hitogarmim v. togarmim 
hitosug, dat., 158 
hontecnatatu, dat., 153 
honuntogaitarni, 147, 160 
hothoil, 165 
huandluithi, dat., 157 
huathati. ace. pi. Jem. v. 

uathath, 149 
huatigitir, 149 

iar, 157, 162 

iarfaichtheo, iarfaigtho, 

dat. /?/., iarfaigthib, 148 
iarfaigid, iarfigid, 148, 165 
iarma, 165 
iarmbaithius, 165 
iarrnai'oicn, 165 
iarm(s)uidigthe, 165 
iarh, 160, 1<>5 
iarsintairgiriu, dat., 153 
iartimmi, 165 
iasg, 163 

icachthir, dat. neut., 157 
id-, 139 
i'/ir, 154, 155 
ldpart, v. edpart, 144 
idpart, 135, 145 
icipart, 151 
il, 163 
ilar, 163 
ildaui; 151 
ilde, 163 
illestur, 159 
iltolmdden, 151 
////, 142 
imbed, 143 

imbradud, iuiradud, 148 
imbrati, imiati, nom. pi., 

ace. iniratiil, 148 
iracabthi, 147 
inicasti, 147 
imchulmritlg, 159 
imdhiden, note, 150 
imdibthe, 147, note, 147 
imm, 140, 142, note, 142 
immfolung, dat. v. do im. 

immidforling, 159 
immumf'orhng, 159 
iuiradud, v. iihbradud, 148 
iruratiu, ace. v. imbrati, 

in', 165 
in, 140, 165 
in, 140, 157, 160, 162 
in-, 140,160 
inchamthuisil, 157 
incholnigud, gen. inehol- 

niehtho, incholnigthea 

inchosc, note, 151, 165 
inchrumn, gen., 157 
inchuimriug, 159 
incomscribhdalth, 151 
ind, 140 
ind', 165 

ind-, 143, 153, 155 
indaerehoiltea, gen, v. 

erchoiliud, 148 
indala, 156 

indegaid-ii, 143 

indibilsigthe, 148 

indi'olaid, gen., 157 

indidultaigae, 150 

indium m, 143, 165 

indtogas, 153 

indtuigther, 147 

induini, 153 

inghean, 159 

ingiun, 165 

ingor, 143 

ingrentid, 149 

inn-, 155 

inna, 157 

unite, 141 

inse, 160 

insenduine, gen. intsen- 

dumi, 153 
insin, 153,160 
inso, 153 
inspirto, gen. 153 
hit, 140, 144 
int-, 144, 153, 155 
intaidlich, gen, 153 
intalrmchrutto, gen. 153 
intesa, gen. 153 
intinnscana, 146 
intisiu, 160 
intithall, 160 
intiu, 141 
inte, 141 
intoichther, 147 
intonnaiglm, 151 
intsamail, 144 
intsamuil, 146 
intsechtaigtha, 148 
intsliucht, 144, 146 
intuisil, nom. 153 
intursitib, 151 
iodhbairt, 145 
iomad, 143 
ion-, 140 
ionam, 143 
iongnadfi, gen. ionganta, 

ionnta, 141 
ir-, 139, 164 
lrchoiltith, 149 
irchollud, 149 
ire, 164 
ireiu, 164 
w, 153,156 
isin, 156 

isindanmaim, 157 
isindepistil, dat. fern. 157 
isinoinchorp, dat. masc. 

isintuisiulsin, dat. 153 
isintuaichli, dat. 153 


Celtic Index to Phonology in Irish. 

is&iri, dat. 159 
Issintodochidui, dat. 153 
itarscarthar, 147 
itchethir, 157 
itchoimthecht, 159 
ith-, 139 
ithim, 162 
itossiich, 165 
itsenmathir, 159 

la, 160 
]aim, dat. 157 
Ian, note 144, 163 
lanad, 163 
lane, 163 
lani, dat. 159 
laur, Icur, 163 
lethan, 163 
Mne, pi. leinte, 141 
lia, 163 
linad, 163, 
linn, pi. linnte, 148 
lintldi, 149 
lobrigthir, 147 
loiscthe, 147 
loth, 139 

mac, 145, 151, 159 
mace, 145, 161 
macthire, 150 
mactire, 150, 151 
maer, M. I., 140 
maini, pi., 140 
mamistrech, M. I., 139 
maitk, 140 
maldacht, 143 
mallacht, 143 
manireltar, 147 
mamidubfeil, 160 
maoin, 140 
maor, 140 
mara, M. I., 139 
marbh, 143 
marc, 145 
nsarafeste, marrufeste, 

niedontairismid, 151 
meite, gen. fern., 157 
messe, 160 
mi-, 151 
miastar, 147 
mile, 153 
mili, 158 
mistae 149 
mo, 157, 159 
mo, 159 
mocbland, 159 
inbga, 139 
moin, pi. mointe, 148 
monistre, 139 

mora, 139 

mora, 157 

mordha, 149 

mothol, 159 

muid, 159 

muintir, muintear, 145 

nrantar, 150 

muntir, dat., 159 

muntith, 149 

munud, 149 

mur, pi. murtha, 149 

na, 155 

nach. 156, 161 

nachibfel, 160 

nadipru, nadipro, 137 

nam-ball, 143 

nammalL 143 

naoih, 142 

na ructhae, v. ructhae 

ndeaghaidk, 143 

neal, pi. ne'alta, 148 

neam, neim, 142 

neart, 145 

ne'b-, 142 

nebmarbtu, nebmaxbtath, 
note 149 

nech, 140, 161 

necht, 161 

neim; 142 

nemed, 139 

neoch, noch, 140 

nephpiandatu, 157 

nert, 139 

nertad, gen. nerta, 148 

nerutsu, 152 

m, 160 

nibtha, 160 

nidan, 160 

nifiastar, 147 

nfgette, 147 

nilfolad, ace, 158 

nimebaratsa, 159 

nimdibi, 158 
nimptha, 159 
nimtba, 159 
m, 160 

nmforteit-nij 160 
ninta, 160 
niscartha, 147 
nisfetemmar, 160 
nfsfitir, 160 
nistabur, 160 
oitenat, 160 
no, 160 . 
nobcarad, 160 
nbbs6irfa-si, 160 
nocretim, 160 
nochrochte, 147 

nodascara, 160 
not, 158 
noib, 142 
nolintae, 147 
nomglantar, 147 
nomthacbthar, 147, 148, 

nonchretid-si, 160 
nondubcairim-se, 160 
nondobsommfgetar, 160 
nondasoirfea, 160 
nongabthe, 147 
nonlmtarni, 147 
nonnertarni, 147 
nonsoirfea, 160 
nopredebhn-se, 160 
nosmoidet, 160 
nosnguid-som, 160 
midanichrocba, 159 
nunsluinfemni, 160 

o, 151, 159 

6', 157, 158, 165 

0'Briain,gen. I. Bhriain, 

dat. d'ua Bhriain, 151 
ocht, 139, 158 
ocht, 135, 139 
od-, 137 
6en, 157, 158 
oen, v. oin, 140, 141 
oena, 157 
oenchoioidiu, 157 
benchorp, 157 
oenchrann, 157 
oile, v. eile, 140 
oin, 140 
oinaichthir, 147 
oinchorp, 157 
oh, 145 

olsodaan, olsodm, 160 
on, 157 
onach, 157 
ond, 159 

ondoentoisrinn, dat., 157 
ontechtairiu, dat., 153 
ontrediu, dat., 153 
orcaid, 139 
orcas, 139 
drdha, 149 
orm, note 137 
orra, ortlia, 137, 141 
orraibh, note 137, 141 
orrainn, note 137, 138. 141 
ort note 137 
ortha, 141 
6-midi, dat. fern, olsulde, 

othad, 139 
oua, v. ua, 141 

Celtic Index to Phonology in Irish. 


peccad^en. pechtha, decl. 

148,157,159, 162 
persan, 162 
persin, 156, 157, 159 
pbersm, 157 
precept, 162 
prim, 162 

rad ride, 143 

rainn, 159 

rancatar, M. I., 145 

rdngas, 145 

re, 162, 164 

recaeb, 157, 165 

reelittaircid, 151 

rect, dat, 159 

relath, relad, gen. relto, 

rem, 142 
rem', 165 
remeperthe, 147 
remfoiti, 147 
rerni, 140, 164, 165 
remib, 165 
remiepur, 165 
remthechtas, 165 
ren, 164 
reri, 160, 165 
renairite, 165 
ri, 137 

riagoil, dat. 157 
rii>hduinte, 149 
ro, 160, 163 
robcar-si, 160 
robclandad, 160 
roberrthe, 147 
robfothiged, 160 
rodcbursach, 159 
rofestar, 147 
rohh, 142 
roimsi, 153 
roime t 141 

roimpe, 140, 141, 153 
roimsi, 141 
rolaumur, 139 
rolin, 163 
rolSmor, 139 
roraam, 140 
romp a, 141 
romsoirsa, 159 
rondasaibset, 160 
rondobcarsam-ni, 160 
rondpromsom, 159 
ronfitid-ni, 160 
ronsoir, 160 
ronsoir-ni, 160 
xoscarsam, 151 
roscomal, 160 
rospredach, 160 
rest an, 151 

rotchechladar, 159 
ructhae, 147 

sa, 153, 160 

saib, saeb, soib, soeb, 140 

saide, saidai, 160 

saigul, 140 

sain, nom. pi. saini, dat. 

sainib, 156 
san, 156 
Balm, 162 
samaltir, 147 
saobh, 140 
saegkal, 140, 156 
saoghalta, 149 
sastai, 147 
scote, scotae, 149 
scothj 149 
scrfbthar, 147 
*se, 141 
se, 153, 160 
se'j 158 
seacht, 158 
seanchus, 156 
seanduine y 156 
sech, 161 
sechim, 161 
sechitir, 161 
secht, 161 
sechtmaine, 161 
sechtaigud, 148 
seim, dat., 157 
sem, 160 
sendulne, 151 
sens, icaeh-, 157 
seutinni, 151 
sedl, pi. sedlta, 148 
serbe, dat. 159 
sgeal, pi. sge'alta, 148 
side, 160 
sin, 153, 160 
siu, dat, 160 
*siu, 141 
siabrad, 149 
slabratae, 149 
sluintir, 147 
so, 153, 160 
sodam, 160 
soib, v. saib, 
som, 160 
soscele, 151 
sosiith, dat. masc, 157 
srathatath, -tat, 150 
srathath, 150 
sruthy decl. 155 
su, 141, 160 
suidigtbir, 147 
suit, decl. 155, 159 
sum, 158 
sulbamchthe, 147, 148 

sulbairigud, 148 
superlalt, 158 

tabairt, 148 

tabbraim, 137 

tabur, 137 

taid-, 135 

taidmenader, taidminedar, 

note 136 
tain, pi. tdinte, 148 
tair-, 144 
tairci, 144 
*tairmchruthad, gen. tair- 

mchrutto, 148, note, 148 
tairmtbecht, 165 
tairngiri, 159 
taith-, 135, 144 
taithminedar, note, 136 
talmande, 149 
tanaise, 156, dat. neut., 

tanaisiu, 157, dat. Jem., 

tanaisi, 157 
tancamar, 145 
tar, 142 
tarais, 165 
tarbh, 143 
tarcrach, 165 
tarmi, 165 
tar(s), 141, 165 
tarsin-, 165 

teanga, pi. teangtha, 149 
tecnate, 149 
teinne, pi. teinnte, 148, 

tenat, 160 
tened, 159 
tenge, 162 
ter-, 144 
tes-, 144 
tesbaid, ace, tesbaitb,c?a^ 

note, 143 
tesst, 146 
tbaidbse, dat., 158 
thdirse, thdirsi, 141 
thdngas, 145 
thdrsa, 141 
thdrsta, 141 

thogbliail, v. dothog, 136 
thoil, dat., 158 
thoisig, dat., 157 
tboi, 157, 158 
thorm, 142 
thor rainn, dat. thorraibh, 

threana, nom. pi. 157 
thrium, gen. masc, 157, 

thual, 159' 
thuare, 157 
thuisiul, dat. mate, 157 


Celtic Index to Phonology in Irish. 

ti-. 135 : tu- 5 144 

il , 141 I tuaithe, 159 

tiar-. 144 tuar-. 144 

tibradaibh,tipradaibh,153 tuati, 149 
tid-, 135 | tuath, 149 

tidbarid, 135 tuisel, 139 

«wn- 3 144 tur-, 144 

timni- 3 144 tussu, 160 

timmorte. 147, 148, 149 
tiranais, M. I., note, 136 
tirtine. note 13-5 
tin-, 144 
tind'-, 144 
tinttith, 151 
tionnsgnadh, gen. tionns 

go.nta, 148, 149 
tipradaibh. v. tibradaibh 

tiprait, M. L, 135 

tapra, tipru, gen. tiprat, uasta, 141 

; ua, v. oua, 141 

■ uad. 165 

: uadfiabcbtbi, 165 

uailbe, v. uall 

uaim, 142 

uaiinru, 142, note 142 
: uair, decl., 155 

uaiste, 141 

uaithe, 141 

uall, ^e/?. uailbe, 143 

tir, 159 
tirthat. 1 50 
to-, 144 
tobar, 135 
tocbaimm, 145 
tog, 136, 137 
togarmim. dat., 157 
togbbaidb, 136 
togbbaiin, 145 
toirthich, toirtbig, gen.. 

tolrse, 159 

toisicb, dat. fern., 157 
tol, 158 
tor-, tor-, 144 
toradh, pi. toirthe, 149 
torbe, 158 
torunn, 141 
tre'ana, v. thre'ana, 157 
treasa, 157 
trebaire, 158 
tremdirgedar, 165 
trenii, 165 
tresin-. 165 
tresinfuil, 160 
tri, 160, 162 
tri, 158 

Irian (d-triari), 158 
trieha, nom. pi. tricbit. 

tricbretim, 160 
triit, 165 
triocka,gen. trio chad, nom. 

pi. triochaid, 156 
triotha. 141 
tris, 165 
trithe, 141 
trithernel, 160. 165 
troeaire. 145 

uatha, 141 

uatbaib,oto. pi. v. uatbatb, 

uatbate, 149 

uatbatb, uatbad, 139, 149 
vghdar. 141, 145 
uile, 156 
uilib, 165, 142 
uimb'si, 153 
uimpe, 141, 153 
uirre. uirri. note 137, 141 
urn, 140, 142 
v. mam, 140 
umpa, 141, 145 
urdhairc,\. eardhairc .note 



adaned (plur.). 164 
adar, r. atar, 164 
aetinet. 164 
abanaff, v. obonaf, note 

atar, v. adar. 164 
bwystuil, 135 
cant-, 144 
e-. 139. 

ebawl, 139. 161 
ed-, 139 

ederyn. v. eterinn, 164 
enw, note, 142 
eterinn, v. ederyn. 164 
etncoilbaam. 164 

im v. ym. note 143 
laws, 163 
liaus, 163 

Uaiver, 163 

llawn, 163 

lied, v. llyd (llet\ 163 

lledanu, 163 

Scam*", 163 

Warns, 163 

?/b»etf, 163 

lluossyd, 163 

llyd, ». lied, 163 

neuat. neuad, 139, 

nertb, 139 

obouaf, v. abanaff, note 

orgiat, 139 
ucher, 161 
uitb, v. wvth, 139 
On, 141 

vrytb, v. uith, 139 
J-, 139 
yd-, 139 
yni-, v. im, note 143 


ar-,139, 164 
at-, 139 
*ep, 161 
er. 164 
et-, 139 
gafar. 162 
gu, 162 
guo, 162 
guor, 162, 163 
gurtb, 162 
bep, 161 
kafael, 162 
laun, 163 
litan, 163 
llydan, 163 
map. 161 
nop, 161 
paup, 161 
penn, 161 
petguar, 161 
pimp, 161 
pise, v. pysg, 163 
plant, 161 
pren, 161 
prenu, 161 
pysg, v. pise 
rac-, 164 
yr, 164 


ebol, 139 

enef, eneff, note 142 
escop, 13 5 
gwesper, 116 

Celtic Index to Phonology in Irish 


(h)anow, note, 142 
hethen, 164 
idne, 164 
leas, 163 
len, leun, 163 
llewer, 163 
loar, 163 
loure, 163 
luas, 163 
nerth, 139 

thym, note 143 
war, 164 

y-, 139 

ydnic, 164 


arvorek (Breton), 165 
dif, diff, note 143 

e-, 139 

enef, eneff, note 142. 

ez-, 139 

gousper," 1Q1 

hanu, note 142 

leun, 163 

nerz, 139 

war t 164 


[The following have been noticed in preparing the Indices Verborarn.] 

Page 102 6, line 13 from the bottom, P. or. should beP. or. 

„ 110 a, line 2 from the top, there should not be a full point after cam. 

„ 121, line 8 from the bottom, for the reference (I. 177, 180) read (pp. 74, 

,, 126, line 4 from the bottom, for noun € } si, ed, etc., read nom. €, etc* 
„ 141, line 9 from the top, for the reference, p. 119, read p. 127. 
„ 158, line 12 from the bottom, for trochaid read triochaid. 
„ 158, line 4 from the bottom, for toll read tol. 


»*AN 1 

4 ?flfl# 

■f a+jwq 



3 9031 01161863 4 


Bapst Library 

Boston College 

Chestnut Hill 67, Mass.